Full text of "On the Osteology of the Hyopotamidae"

See other formats

L 19 I

I*. On the Osteology of the Hyopotamidse. By Br. W. Kowalevsky.

Communicated by Professor Huxley, Sec. M.S.

Beceived December 19, 1872, — Read February 6, 1873.

Introduction.

The paper which I lay before the Society is an attempt to treat with sufficient osteo-
logical detail an extinct family of Ungulates which had an immense range of distri-
bution and a great variety of forms in the two periods of the earth's history which
preceded our own. The fate this family has met with at the hands of palaeontologists
is a somewhat sad one, presenting a warning example of the unscientific method that
was paramount in the palaeontology of the Mammalia after the time of Cuviee. With
the exception of England, where the study of fossil Mammalia was founded on a sound
basis, and some glorious exceptions on the continent, we have very few good palseonto-
logical memoirs in which the osteology of extinct mammals has been treated with
sufficient detail and discrimination; and things have come to such a pass, that we
know far better the osteology of South American, Australian, and Asiatic genera
of fossil mammals than of those found in Europe. Nearly all fossil Mammalia
which have been described in detail belong to genera that still exist on our globe, or
whose differences from fossil forms are trifling. After the splendid osteological investi-
gations of Cuviee, had revealed to science a glimpse of a new mammalian world of
wonderful richness, his successors have been bent rather on multiplying the diversity of
this extinct creation, than on diligently studying the organization of the fossil forms
that successively turned up under the zeal of amateurs and collectors.

From the year 1828, and even before, when Laizee, Pomel, Ceoiset, and others began to
give short notices on the Mammalia of Auvergne, mammalian genera and species from this
locality have been multiplied at a prodigious rate, every private collector giving his own
generic and specific names, with no better description than stating the real or supposed
number of teeth, and some phrases as to the general resemblances of the fossil in
question. Others substituted in their short notices other names, while the scientific
work of description did not proceed further than the mere counting of the number of
teeth. This process has given rise to such an utter confusion in the palaeontology of
the extinct Paridigitata, that even now (forty years after the date of the earliest
notices) we are utterly ignorant of the true extent and organization of the Miocene
mammalian fauna of Auvergne, for instance — though materials for a detailed study of
the subject abound in all great public, and many private, collections, the fossils being
very common. No palaeontologist, even of the highest standing, could boast of knowing,

MDCCCLXXI1I. E

20 BE. W. KOWALEVSKY ON THE

in our own time, what Dremothermm, Dorcatheriwn, Elapliotherium, Gelocus, and so on
really are, what are the bones belonging to each set of teeth (as the names were mostly
given to these last), whether they had horns or were hornless like the Tragididce^ and
so on. If we add that German authors described the genera of Paridigitates which were
found and named in France under different names (as Palceomeryx^ Microtherium,
Hyotliermrri) and so on), when they came from German localities, the confusion may
be guessed. Having no good descriptions and no figures of the genera noticed in France,
the German authors almost necessarily fell into the mistake of renaming what was
already named. Once named, the genus was allowed to go forth with the short and
wholly insufficient characteristics given to it by the first describer, the impossibility of
adding one's name after the generic or specific designation seeming to take all interest
from it. And this, moreover, is the best case; for frequently the same form was>
described by another palaeontologist under a different generic name, or, if this was-
utterly impossible, a new species was made of it, founded on some difference in size or
other trifling character. Happily, however, a reaction began to set in, one of the first
to head it on the Continent being Rutimeyer, who did not confine his study merely to
the teeth of fossil Mammalia, but aimed with brilliant success at a complete investi-
gation of the osteology of the extinct genera and of their affinities with the living
ones. Gaudry's work on the fossils of Pikermi (the best palseontological work that has
appeared in France since Cuvier's 'Ossemens Fossiles'), Fraas's ' Fauna von Stein-
heim,' Alphojntse Milne-Ed wards's ' Oiseaux Fossiles,' and many others may be cited
as examples to prove that the new tendency has fairly set in and will bear good fruit.
The wide acceptance by thinking naturalists of Darwin's theory has given a new life
to palaeontological research ; the investigation of fossil forms has been elevated from a
merely inquisitive study of what were deemed to be arbitrary acts of creation to a
deep scientific investigation of forms allied naturally and in direct connexion with those
now peopling the globe, and the knowledge of which will remain imperfect and incom-
plete without a thorough knowledge of all the forms that have preceded them in the
past history of our globe.

The foregoing observations are intended only as a sort of apology for the somewhat
minute osteological details into which it seemed to me necessary to enter in my descrip-
tion of the two genera which form the subject of the present memoir ; before, however,
we proceed to the concrete description of their skeleton, it is necessary to offer a few
remarks on the position they hold among other fossil Paridigitata, as it seems to me
that it has not been duly recognized by any previous author.

In, all our speculations about the history and origin of the Paridigitata, a paramount
importance has always been ascribed to Anoplotherium^ as the most ancient form of
the Paridigitate series. Now, seeing the reduced state of the skeleton of Anoplotlieriimi,
there cannot be the slightest doubt that this position is an entirely usurped one. How
this state of things originated is easily accounted for. Anoplotherium had the good
fortune to be found and described by Ctjvier, who gave a thoroughly good description

OSTEOLOGY OF THE HYOPQIAMIDiE. 21

of its osteology, and, comparing it with some of the living Paridigitates, found that it
and they possess many features in common. The naturalists who have ventured to
theorize during the period of purely descriptive science which lies between the time of
Cuvier and the complete revolution caused by Darwin's great work, looked always at
the accurate and talented description of Cuvier, finding no good materials in the works
of modern palaeontologists from which to draw their means of generalization. As this
absence of detailed descriptions of fossil European Mammalia was prolonged till our
own time, we can understand how, even after the great revolution caused by the publi-
cation of Darwin's views, all writers that were leading the new movement of evolution,
and trying to apply the theory of descent to our modern Paridigitata, had still to consult
the works of Cuvier to find full and accurate information. There they could find only
Anoplotherium fully described (as the Xvphodonand, especially, Dichohune were very much
neglected in the ' Ossemens Fossiles ') ; and as it was, moreover, the most ancient form
known, they placed it at the commencement of their pedigrees of Paridigitata. But if
we consider the structure of the feet in Anoplotherium, we cannot avoid the conclusion
that this genus is only an exceedingly reduced form, descended from some earlier
Ungulate of the early Tertiary or, more probably, Cretaceous period. Anoplotherium is
clearly the last remnant of a dying-out branch, in no case the progenitor of the wonder-
fully rich and diversified Paridigitata which succeeded it in the Miocene period, and
which became so enormously developed in the Upper Miocene and Pliocene epochs, form-
ing in our own time perhaps one of the most richly developed of animal groups. The
feet of Anoplotherium are so much reduced, presenting only two developed meta-
carpals and metatarsals, with merely rudiments of the lateral toes, which certainly existed
in its ancestors, that we cannot imagine such a reduced form giving rise to Miocene or
even, modern Paridigitata, many of which have four completely developed metacarpals
and metatarsals. Besides, it is a very general truth that only those families which
were exuberantly developed in bygone times, presenting many subgenera and a great
variety of specific forms of different size, have had any chance of leaving a progeny
behind them. "We see examples of this in some of our recent genera the pedigree
of which is now very completely known. There can be, in my opinion, no reasonable
doubt that the Horse descended from the Palmotherium (very probably the Ehinoceros
had the same origin from a Palseotheroid form, though this is not so certain); and see
what immense diversity we find in the Palceotheridce of the Eocene and Miocene epochs.
The quantity of described species of Palceotherium is only a small fraction of the
quantity that really existed, as every one who looks through a large collection of
Eocene teeth becomes aware. Besides, the Palwotheridw range in size from P. mini-
mum (known only by its metapodium), not larger than a rabbit, through all inter-
mediate sizes to P. magnum, fully as large as our Rhinoceros. We may mention that
the Paloplotherium and Anchilophus belonged also to this group, and are extensively
developed in the Eocene epoch. Only such prolific types, sending branches in all direc-
tions, have any chance of not wholly dying out in the course of time. If, in the struggle

22 DE. W. KOWALEVSKT ON THE

for existence, through geological changes of climate, slow submergence of continents,
and elevation of the former sea-bottom to the height of the Himalayas, many genera
must have been destroyed, still some one branch may have remained, and by gradual
modification through natural selection, and perhaps by the agency of some other un-
known cause, has given rise to new genera and species better fitted for the changed
circumstances of life.

We see the same truth illustrated in the case of the Tapir, which is the last repre-
sentative of a group extensively developed in the Eocene epoch, as all those genera
known under the names of Pachynolophus, Lophiodon, Hyracothermm, Pliolophus be-
longed clearly to the family of Tapirinw.

On the contrary, Anoplotherium is exceedingly poor in specific forms, only two being
known with certainty, without any great difference of size ; and though future discoveries
may possibly increase the number of species, it is doubtful if ever they were as numerous
as those of other genera which peopled the earth at the same period. In my opinion,
as I shall try to prove further on, the Anoplotherium is an aberrant and very reduced
branch of the early Eocene Paridigitata, which has no direct connexion with the living
ones, and the true line of descent of our Ruminantia must be traced through other
genera of the early Eocene epoch.

If, on the other hand, we turn our attention to the Ilyojpotamidce, it must be con-
fessed that the richness of this family in subgeneric and specific forms ranging through
all sizes is really astonishing; and we shall be hardly guilty of exaggeration in saying
that the diversity of Hyopotamidce in the Eocene and Miocene times was as great as the
wonderful diversity of the Euminantia is at the present day. That this is not fully
acknowledged by palaeontologists is due to the neglect with which this family has been
treated, and which I shall endeavour to repair.

Under the name of Hyopotamidw I understand all the Eocene and Miocene Paridigi-
tata which had crescentic teeth, with five well-developed lobes on their upper molars.
The family might be termed Anthracotheridw, as the Anthracotheria are among the
most prominent representatives of this group; but the Hyopotamidw being richer in
subgeneric forms, the family may perhaps better bear their name.

The chief character on which the genus Hyopotamus (under the different names of
Ancodus, Cyclognathus, Bothriodon) was founded is the shape of the upper molars, of
greater breadth than length, and having five well-developed crescents or lobes* No
author has mentioned any bones belonging to different sets of teeth, on the real or
supposed differences of which many species were founded. The priority of mentioning
these teeth under the name of Ancodus is claimed by Pomel in his 4 Catalogue ; ' but
as he neither gave a good description of them, nor illustrated his short notices by figures,
no palaeontologist has accepted this name, and it may be considered extinct. In fact,
long before Pomel, Cuvier, having received a jaw of this genus from Puy, compared it
with Chceropotamus and Anthracotherium ; and as he had only the back molars, he
deemed it unnecessary to separate the new animal from Anthracotherium, and designated

OSTEOLOGY OF THE HYOPOTAMIMK. 23

it under the name of Antliracotherium velaunum. In the year 1847, however, Professor
Owen, having received from the Marchioness of Hastings very good materials, as far as
the dentition was concerned, studied with great accuracy the structure of these teeth,
and found in them sufficient characteristic differences to justify their separation from
Antliracotherium under the name of ffyopotamus. Notwithstanding that one of the
characters on which the distinction was based (the complexity of the upper premolars)
was found to be a mistake, these complex teeth being milk-molars and not premolars
of the permanent dentition (see Quart, Journ. Geol. Soc. vol. iv. 1848, pL vii. fig. 6 ; ]f
and p 3 are in reality d* and $ 3 ), still there were other characters which entirely justified
the distinction, as I shall show when we arrive at the description of the teeth, Pro-
fessor Owen gave such a thoroughly good and accurate description of these teeth,
accompanied with plates, that he may be considered the founder of the genus. One
year later, M. Aymard, of Puy, in a footnote to his paper on Entelodon, mentioned the
Hyopotamidw from Eonzon (Puy) under the name of Bothriodon, as he believed them
different from those described by Professor Owen from Hempstead. But having com-
pared an extensive series of jaws and paijjts of skulls from both localities, I have been
unable to detect any difference in them, and therefore the fossils from Puy must be
included under the same generic name as those from Hempstead. The only author
who has accepted the genus Bothriodon of Aymard is Professor Gervais, who, in his
6 Paleontologie Fran§aise,' p. 192, says that the Ilyopotamus is to be distinguished
from Bothriodon ; but I have not been able to find any reasons adduced in the de-
scriptions of the two genera why Professor Gervais considers them distinct, or on what
characters he has founded his distinction. Notwithstanding the care with which I insti-
tuted my comparison, I could not detect any differences between the Hyopotamidw from
Puy and those from Hempstead; and, so far as I can see, their separation into two
distinct genera is entirely unfounded. Professor Gervais has described many portions of
jaws from different localities in France, and has given them different specific names, the
distinction being founded on size and on the fact of their having been found in different
localities. As I reserve the discussion of specific differences for the end of my paper,
I will not enter into the criticism of these doubtful species just now.

In the year 1861, Professor Kutimeyer figured and described, with great accuracy,
some few upper molars of a small Hyopotamoid animal from Egerkingen. These teeth
had, however, been previously referred by H. v. Meyer to a [new genus, Tapino&on.
Professor Butimeyer had only true molar teeth from Egerkingen, and as they really did
not present any fundamental difference from the Hyopotami of Professor Owen, he
considered them to belong to the same genus. The Egerkingen specimens being very
small, Rutimeyer separated them specifically under the name of Hyopotamus GresslyL
This discovery was very important, as it carried the genus into undoubted and even early
Eocene strata*. Then came the discovery of the wonderfully rich Eocene fauna of

* I have no doubt that the small Eocene Byopotamidce -will one day be generically separated from the true
Hyopotami, as known from Puy and Hempstead, as their premolars are somewhat different. I shall not do

24 DR. W. KOWALEYSKY ON THE

Mauremont, described in two memoirs by Pictet, De la Harpe, and Humbert. Here
the Hyopotamidm were much more numerous than at Egerkingen, and presented a great
variety of size. Among these remains were some specimens identical with those found
at Egerkingen, and they were correctly referred to the Hyopotamus GresslyL The
largest species was said to be identical with Hyopotamus crispus, Gerv. ; but as Professor
Geryais, in the second edition of his ' Paleontologie Frangaise,' referred the Hyo-
potamus crispus to Xiphodon (without any suitable ground, as it seems to me), the
whole matter must be revised once more. I cannot refrain from stating here that, in
their second memoir, Messrs. Pictet and Humbert have mixed jaws belonging to different
animals in a very strange manner. For instance, the upper jaw figured (I c. plate xxiv.
fig. 3, c) as Bhagatherium is in reality a Hyopotamus, and the lower jaw (fig. 2) bears not
the least resemblance to the true Bhagatherium mandible figured in the first memoir,
and, in my opinion, belongs to Hyracotherium or Anchilophus. In the same paper,
moreover, the two authors have figured and described a very interesting small Paridigitate
mammal, which they called Cainotherium Benevieri; but there is not the slightest doubt
that this small Ungulate cannot be referred 1*o the genus Cainotherium. I have stated
before that the chief characteristic distinction of the molars of Hyopotamus, as described
by Professor Owen, consists in having five distinctly developed lobes or crescents to their
upper molars. These five lobes or crescents are disposed transversely, three on the
anterior half and two on the posterior half of each upper molar, as may be distinctly
seen in Plate XXXIX. fig. 1. With the exception of Bichodon and Merycop.otamus,
all Eocene and Miocene Paridigitata with crescentic teeth had always five lobes on their
upper molars, disposed in the same way as in Hyopotamus ; and, so far as we know at
present, the position of the five cusps of the upper molars is reversed only in two very
characteristic genera, two being placed in front and three behind. These two genera
are Bichohune and Cainotherium, which also by their osteological characters seem to
stand in a very near and direct relation. Every mammalogist is aware how constant are
the dental characters in large groups of Mammalia ; and if I state that such different
genera as Camelopardalis, Camelus, Cervus, and Bos show less amount of difference in the
structure of their upper molars from one another'* than exists between the molars of
Cainotherium Benevieri and the Cainotheria from Auvergne, every one will readily admit
that the so-called Cainotherium Benevieri, in which the five cusps of the upper molars
are disposed in the same way as in all Hyopotamoids (three in front and two behind),
cannot be put into the same genus with the true Cainotheria, in which the disposition
of the cusps is reversed — two in front and three behind. (Messrs. Pictet and
Humbert noticed this difference in their description, but they did not consider it im-
portant enough for a generic distinction.) The upper premolars show also some

this in the present paper, as I hope to collect more ample materials, not only for the dentition, hut also the
skeleton of the Eocene Hyopoiamidce.

* Professor Owe]* i n Quart. Journ. Geol. Soc. vol. iv. p. 111.

OSTEOLOGY OE THE HYOPOTAMID^E. 25

differences froift the true Gainotheria, and the lower molars even more. If we consider
each lower molar as composed of four parts, two crescents on the outer side and two pillars
closing the crescents on the inner side (Plate XXXIX. figs. 8-12, ac,pc, ap,pp), then we
may state that, in the true Cainotheria from Auvergne, the back part of each posterior
inner pillar is a little prolonged backwards into a small additional cusp ; this prolonga-
tion is especially marked on the inner side of the last inferior molar, making the pos-
terior additional talon of this tooth quite double. This difference may be clearly seen
in comparing a lower jaw of a Gainothermm from Auvergne with the enlarged figures of
Pictet ('Faune siderolitique,' plate xxvi. fig. 9, c). By all these characters the Caino-
therium Benevieri, Pict., differs from the true Cainotheria, and agrees entirely with
the Eocene Hyopotamidce figured by Rutimeyee and by Pictet himself. Therefore the
Gainothermm Benevieri cannot remain in the genus in which it was put by Pictet,
but must be arranged with the rest of the Eocene Hyopotamidw, as Hyopotamus
Benevieri*.

I cannot refrain from stating that, in my opinion, the five-lobed character of the upper
molars is of a too general value to be used for generic distinction. In fact all the
Eocene and Miocene Paridigitate genera (with crescentic teeth) have five-lobed upper
molars f , and therefore this character is as unfit for generic distinction, in this large
assemblage of animals, as the fact of having four-lobed molars would be found unfit if
we tried to apply it to the living Euminantia ; it is of too general a nature, all living
Ruminantia having four-lobed upper molars.

In such cases where the true molars present too great a uniformity for furnishing
good distinctive characters, the shape of the premolars may be of great use, as was well
shown by Professors Rutimeyee and Hensel, in reference to Suina and Ruminantia.
And, in fact, if we compare the premolars of the Eocene Eyojpotamidw figured by Pictet
and Humbeet, Hyopotamus Oresslyi, E. Benevieri, and even his H. erispus (I. c. plate
xxiv. fig. 11), we shall see that all these Eocene Hyopotamidw, though of such different
size, agree together in the shape of their premolars, while they differ by the same cha-
racter from the true Miocene Eyopotamidm from Hempstead and Puy. As I hope soon
to collect materials for a description not only of the dentition but even of the skeleton
of these Eocene species, I will not enter further into this matter here, and I will only
state that, by comparing the figures of Pictet with my plates, the reader will perceive
that the premolars of all the Eocene Hyopotamidw are, so to say, more ruminant-like :
this is especially the case with premolars jp 2 and p s ; they are considerably more elongated
and not so high as in the Hyopotamus and Anthraeotherium. Resides, 1 suspect from
some bones seen in the collection in Lausanne, and especially from a metatarsal of
Hyopotamus (Cainotherium) Benevieri figured by Pictet (I. c. plate xxvii. fig. 2), that
some, if not all, of the Eocene Hyojpotamidw were diclactyle, at least the metatarsal

* The Eocene Cainotherium Courtoisi from Yaucluse is identical with O, Benevieri, and therefore must
share its' fate, and he united to the Hyopotamidce*
f Except Dichodon and Merycopotamiis.

26 DR. W. KOWALEVSKY ON THE

figured by Pictet belonged undoubtedly to an Ungulate with only two metatarsals.
At all events, these Eocene Hyopotamidce seem to form a separate group ; and as soon
as their organization is better known, they will certainly be generically separated from
the true Hyopotami. As the matter now stands, however, this Hyopotamus (Cainothe-
rium) Benevieri is the smallest representative of the family, being hardly larger than a
good-sized rat ; the Hyopotamus Gresslyi had perhaps the size of our recent Tragulidce.
Between these small Hyopotami and the largest Hyopotamus bovinus, Ow., we have all
the intermediate sizes distinguished as separate species\mder the names of Hyopotamus
porcinuS) Gerv., crispus, borbonicus^ velaunus, Aym., leptorhynchus^ platyrhynchus, vec-
tianus 3 Ow,, and finally bovinus. All these numerous species were, for the most
part, founded merely on real or supposed differences in the size of the true molars,
no author having figured or described any other part of the skeleton. Now, as every
naturalist is aware, there can be no worse method of specific distinction than size,
especially in richly developed families. Take our living Ruminantia, imagine them
to be extinct, and some future palaeontologist trying to give them generic and
specific names by the characters of their upper and lower molars, I do not think he
could arrive at any thing approaching truth. We may certainly form a large series
of ruminant molars, entirely similar in shape, and ranging in size from the small
Antilope pygmwa to the largest Antelopes and Bovidae*, and then distribute all this
assemblage of teeth into different species according to size; but the scientific value
of such species would be indeed very doubtful. Still this is what we see constantly
done in palaeontology. In my opinion, we have, in most cases, not the least chance of
hitting right in establishing specific distinctions on fossil remains; and most of the
published species of extinct animals are certainly only productions of our pakeonto-
logical methods of inquiry, which had no real distinction whilst living. On the other
hand, it is just possible that the real number of so-called distinct species was greater
than we can distinguish by fossil remains ; at all events we have very little security that
most of our specific distinctions correspond to the real state of things which existed in
past geological time. Seeing the impossibility of arriving at any thing like an accurate
knowledge of the specific distinction of extinct forms, it would be much more profitable
to science if we were to give a pretty large range, as far as size is concerned, and concentrate
all our discriminating powers on such characters as are really organic and fundamental, and
may be taken as a basis for generic or subgeneric distinction. To found a new genus,
a palaeontologist is required to adduce some good reason for doing so ; he is obliged to
point out some organic difference, and this leads to a more complete study of the fossil
forms; while the laxity with which we regard species requires, in a great majority of
eases, no other reason than the phrase " this tooth seems to me to be specifically distinct
from one already described," and a new name is formed which goes to the large
number of others uselessly obstructing the science. Every one who has tried to ascer-

* The upper molars of some Ruminantia sometimes present distinctive characters in the form of additional
pillars, but these are often present or absent in widely different forms.

OSTEOLOGY OP THE HYOPOTAMIM. 27

tain the true organization of extinct types, and who is aware of our ignorance in this
respect, will concur with my statement, that if, from the beginning of mammalian
palaeontology, all specific distinctions had been disregarded, and only the generic forms
had been studied in detail, we should know much more about the extinct creation than
we do now. The family of Hyopotamidce may also serve as an instance of the unscientific
methods which are paramount in palaeontology. During the twenty-five years that the
genus has been introduced to science, we have contrived to make more than ten species
out of it ; while nobody has ever cared to ascertain what its organization really was, and
not a single bone has been figured up to this day.

This was the state in which I found our knowledge respecting Hyojpotamus, when I
determined to study it more completely than had been done heretofore. I knew that
the principal collections, containing large materials for the study of this genus, were
at Puy in Central France and in the British Museum. My kind friend Professor
Gaudry, of Paris, gave me letters of introduction to M. Aymard, who is the possessor
of the largest collection in Puy, and to M. Robert, Director of the public museum of
the same town. I met with the most cordial reception from both these gentlemen;
M. Aymard especially, with the utmost liberality, placed the riches accumulated by
him, during thirty years of diligent collecting, entirely at my disposal, and allowed me
to make casts from every specimen I liked. This permission, which is so seldom
accorded by private collectors, was of immense value to me, as my Plates could not be
drawn at Puy, but had to be made in London. The same facility was afforded me by
M. Eobert ; and I take the opportunity of expressing my warmest thanks to both these
gentlemen. My thanks are also due to the artist, M. Pellegrini, in Puy, who made
my casts, sparing no time and taking much trouble, the specimens from which the casts
were taken being generally exceedingly brittle.

I found in the collection of M. Aymard a large quantity of bones, which enabled me to
get a complete conception of the skeleton of the interesting genus described in this paper.
The bones proved that the Hyopotamus was one of the extinct Paridigitata with cres-
centic teeth, and had four completely developed digits on the fore and hind legs. The
size of the teeth and bones enabled me to distinguish two or three species, as had been
previously supposed by M. Aymard. From Puy I came to London, where, by the kind
permission of Professor Owen and Mr. Waterhouse, the bones of the Hyopotamus con-
tained in the collections of the British Museum were placed at my disposal. The jaws
and bones from Hempstead, part of which were described by Professor Owen in 1848
(Quart. Journ. Geol. Soc), proved to be entirely identical with those found at Puy ; so
that the name of Bothriodon, which was applied to teeth and bones found at Puy, under
the impression that they were distinct from those found at Hempstead, could not be
retained, and the priority remains with the name Hyopotamus given by Professor Owen
to specimens from Hempstead. But, besides the bones from Hempstead, I found in
the British Museum a number of well-preserved long bones, some metacarpals and
metatarsals, as well as a well-preserved tarsus, which came from Hordwell from a true

mdccclxxiii. f

28 Dll. "W. KOWALEVSKY ON THE

Eocene bed, and were supposed to belong to Hyopotamus bovinus, under which deter-
mination they were placed in the Museum. At first I thought that these bones might
really have belonged to the Hyopotamus bovinus, though the British Museum possessed
no teeth of this genus from Hordwell. The long bones from Horclwell were strikingly
similar to the Hyopotamus bones from Hempstead and from Puy ; the calcaneus, astra-
galus, and phalanges were so entirely similar to Hyopotamus bones from these two
localities, that it seemed impossible to doubt their generic identity. What looked very
convincing also was the shape of the last, or ungual, phalanges. These ungual phalanges
are very peculiar in the Hyopotami from Hempstead and from Puy ; they resemble
somewhat a large and thick human nail, and we know not a single living or fossil
Ungulate having such peculiar ungual phalanges ; however, even in this the animal
from Hordwell entirely resembled the Hyopotamus. But the metacarpals and meta-
tarsals of the Hordwell animal proved very different from any thing found at Hempstead
or in Puy ; they were considerably larger, and their shape was altogether different ; the
section was much more round, and the inferior or distal ends proportionately thicker.
At first I could not find any plausible reason why the animal from Hordwell, being so
similar to the Hyopotamus in all the long bones of the skeleton, should present this
very striking difference in the shape of the metacarpals and metatarsals. But gradually,
as I grew more familiar with all the peculiarities of these bones, and after having com-
pared them with a large series of metacarpals and metatarsals of extinct and living
Paridigitata, it became quite clear that the animal from Hordwell, although so similar
in all its long bones with the Hyopotamus from Puy and Hempstead, differed widely
from this genus in having only two metacarpals and metatarsals, and not four — being in
fact didactyle, like the Anoplotherium. As we have never yet found at Hordwell a
complete fore or hind foot in its natural connexion, I was very cautious in drawing my
inferences from scattered and mostly broken bones ; but the study of the relat ions of
the carpal and metacarpal, and especially tarsal and metatarsal bones, gave altogether
the same result ; indeed there was no possible doubt that the animal from Hordwell
could not have more than two complete digits to its fore and hind foot. Unfortunately
I could not find any teeth belonging to the new animal whose almost entire skeleton was
before me. The only authority who has noticed Hyopotamus teeth from a deposit in
Hampshire is Professor P. Gervais. In his ' Paleontologie Fran^aise,' while speaking
of Hyopotamus (p. 191), he tells us of having met Hyopotamus teeth, similar to those
described by Professor Owen, in the collection of Bowerbakk ; " they came from a fresh-
water stratum in Hampshire." As the bed in which the bones of the new didactyle
animal were found at Hordwell has really an outcrop in the New Forest, I suspect that
the teeth seen by M. Geevais really came from this bed. However, the absence of
these teeth, though much to be regretted, is not an obstacle to a complete knowledge
of the new genus. The classification of Paridigitata. is based entirely on osteological
characters; and as nearly all the bones of the new animal from Hordwell are known, it
is perfectly characterized, and, in my opinion, much better than many genera of which

OSTEOLOGY OF THE HYOPOTAMIM. 29

only the dentition is known. Besides, we must not forget that the Hyopotamidce, as
proved by the fauna of Mauremont and Egerkingen, are a true Eocene family ; and the
presence of a large representative of this family in the Upper Eocene of England is a
very natural occurrence. I stated before that, even in the Mauremont and Egerkingen
fauna, some of the Hyopotamidw seem to have been didactyle ; and we may take this
as pretty certain as regards the small Hyopotami^ Benemeri. The new didactyle animal
from Hordwell is only a larger representative of these Eocene didactyle Hyopotamidw ;
and we may expect that when its dentition is completely known it will very probably
resemble that of the Eocene Hyopotami as found at Mauremont and at Egerkingen,
and differ by its more ruminant-like premolars from the true Hyopotami which occur
at Hempstead and in Puy. The presence of didactyle genera in Eocene deposits,
while we find tetradactyle genera belonging to the same family in the Miocene, is no
objection to the theory of descent as it is often argued by the adherents of the
special creation hypothesis. The primary stock was undoubtedly tetra- or even penta-
dactyle; and under the incessant tendency to greater reduction and simplification of the
limbs, which we witness in all Ungulata without exception, there were given off side
branches which reached this utmost reduction of the limbs in the Eocene and became
extinct, while the original unreduced stock continued to live till the Miocene period,
A similar case may be imagined in relation to the recent Suinee. There is no doubt
that the Dicotylidw represent one of the most advanced and reduced branches of the
family of Suina ; they practically reach nearly the same state of reduction of their limbs
as the didactyle genera, their lateral digits being only useless appendages, having no
importance for locomotion. Imagine that, by some geological change, the Dicotyles
should become extinct in South America, while other continents should continue to be
peopled by unreduced typical Suina : in this supposed case we should have an extinction
of the filial branch, while the parent stock would continue to live and flourish. In the
same manner I have little doubt that the didactyle Hyopotamidce found at Hordwell
and Mauremont descended from a tetradactyle stock, which very probably presented the
same structure of the skeleton as we find it in the Miocene Hyopotamidw from Puy and
Hempstead*. And though the didactyle genus is found in strata older than those
which gave us such complete materials for the restoration of the tetradactyle Hyopota-
midee, still, seeing the similarity of their skeleton, we may consider the didactyle genus
from Hordwell as a reduced descendant of a form very similar in its skeleton to the
Hyopotami of Puy and Hempstead.

Having ascertained the existence of this reduced representative of the family of the
Hyopotamidw^ I could not, in view of the difference in the number of digits, permit
the new form to remain in the genus Hyopotamus; and though I strongly object to the
creation of new names, there is no help for it in this case, and a new generic division
must be made to receive the didactyle Hyopotamoid from Hordwell, as well as similar
forms which may turn up in the future. As the chief distinction of this genus is its

* The probability is converted nearly into complete certainty if we consider that the reduced didactyle gemrn
presents rudiments of two additional digits, the second and fifth.

30' DE. W. KOWALEVSKY ON THE

didactylity, I propose to name it Diplopus, and to add to the special form found at Hord-
well the name of M. Aymard, who has contributed so much towards the advance of our
knowledge of this extinct group. I will therefore describe the animal from Hordwell
under the name of Diplopus Aymardi.

From the point of view of pure descriptive osteology, it would be perhaps better to
describe the Hyopotamus from Hempstead and Puy and the Diplopus from Hordwell
separately. I have been prevented, however, by many reasons, from following this course,
and I have preferred to give a comparative description of both genera : as the number of
my Plates was limited, I could not figure all the bones belonging to both; and in
the case of entirely similar bones, I gave preference to those which were better preserved.
The present paper contains only the description of the long bones of the skeleton and
of the limbs. I hope soon to be ready with the description of the vertebrae, skull, and
dental characters. My best thanks are due to the officers connected with the Osteo-
logical and Geological Departments of the British Museum, as every thing which could
in any way favour my studies was accorded in the most liberal way. My special grati-
tude is due to Mr. William Davies, of the Geological Department, to whose kindness
and the interest he has taken in my work I owe very much. I am indebted to him for
many valuable suggestions, and for the unremitting kindness with which he assisted me
in looking over and over again through the rich stores of fossil remains contained in
the galleries of the National Collection, and for aiding me in my comparisons in every

possible way.

The Bones of the Skeleton.

The Scapula (Plate XXXV. fig. 1, f nat. size). — This is a bone that has generally the
least chance of being preserved fossil, owing to its flat shape and consequent thinness.
Happily, however, thanks to the excellent method by which the bones from Hordwell were
collected by Mr. Keeping, we have now in the British Museum, besides several detached
fragments, a complete right scapula* from Hordwell, which belonged to the didactyle
animal named by me Diplopus Aymardi. The chief characteristic of this fossil scapula is
its enormous breadth compared with the antero-posterior length. No living Ungulate
shows us such a broad scapula, the nearest approach to it being made perhaps by the
scapula of Hippopotamus. Among the fossil Ungulata, the Anoplotheriunjb (Blainville,
Ost. Anopl. pi. iii.) comes even nearer to it, by the broad expanse of the horizontal
part and by the large acromial process ; there is, however, a marked difference between
the two in the shape of the glenoid cavity, which is very oval in Anoplotherium and
nearly circular in Diplopus. The coracoid process is not very prominent, but larger
than in the recent Suinse; it is separated from the glenoidal fossa by a slight notch,
retreats a little backwards, and does not reach the level of the inner margin of the
fossa. The fossa glenoidea is moderately deep, with a sharp margin raised all round.

The spine of the scapula is very high and sharp, set very obliquely on the horizontal

* As many bones were drawn without mirror (not reversed), it may happen that some of them described as
right may be really left in the original. I shall, however, adapt my descriptions to the Plates, and describe the
bones as they appear in the drawing.

OSTEOLOGY OF THE HYOPOTAMIMS. 31

part, and inclined to the outer side; its upper margin is rugose, but presents no
flattening or expansion, as in many Ungulates of the Imparidigitate series. The pars
acromialis is produced forwards in the same way as in Anoplotherium, the Camelidce^
and most Kuminants, but never in Pigs, where the spine rises very slowly from the neck
backwards without any trace of an acromial process. The rising of the spine above
the surface in Ilyopotamus commences at a distance of 30 millims. from the glenoidal
border ; but the acromion is so produced forward that it reaches nearly to the level of
the glenoidal cavity, overarching the neck of the bone. The supraspinous is a little
smaller than the infraspinous fossa, but the difference is not important.

In considering the very great breadth of the scapula of Ilyopotamus, we must not
forget that all Ungulata have a very large upper cartilaginous border, which is wanting
in our fossil : the scapula figured by me belonged to a young individual ; and we may
safely infer that by the gradual ossification of the cartilaginous upper margin, the
breadth would not be so disproportionate to its length as it is now. On the outer side
of the neck of this scapula we see a pretty deep and large elongated fossa, found also on
two other broken specimens. In general shape this fossil scapula agrees most with
the scapula of Hippopotamus, by its great breadth and by the production of the acromial
extremity of the spine, in which respect it also closely approaches some ruminants.
With the Suinse, however, we find no relation at all; and I particularly insist on this
point, as, owing to the poorness of the Paridigitate types in our recent fauna, we are
very apt, whilst studying fossil Mammalia with unreduced skeleton, to find resemblances
with Suinse even where they are very slight, just because the pig has one of the most
complete (unreduced) skeletons among the living Ungulata.

Unfortunately I could not find in the collection of M. Aymard, nor in that of the
Puy Museum, a scapula of Hyopotamus.

Dimensions of the Scapula of Diplopus AymardL

millims.

Height of the articular fossa . . 34i|-

Transverse breadth of the articular fossa 37

Height of the spine . 31

Transverse breadth, including the coracoid process . . , 57

Breadth of the neck . . 47

Largest transverse diameter 190

Whole length 218

This was written and my Plates were drawn in London in the summer ; but having in
November last paid another visit to Puy, in order to examine the collection of Mr. Vinay,
which I was prevented from seeing on my first visit, I found there a very good speci-
men of the scapula of Ilyopotamus. The upper margin of the bone was broken, and
only about two thirds remained entire; the spine of the scapula was preserved, although its
acromial part was broken; and I could not ascertain if it was prolonged forwards in

32 DE. W. KOWALEYSKY ON THE

the same manner as in the scapula of Diplopus. My paper and Plates being already
finished, I could not give a figure of this new specimen, and shall try to supply the
want of it by a few explanatory remarks.

The scapula in the collection of Mr. Vinay belonged evidently to the largest species
of Hyopotamus found at Puy, and in size equalled the scapula of Diplopus figured
in Plate XXXV. The general aspect of this new specimen presented a great simi-
larity to the one figured from Hordwell ; beginning from the neck, the bone broadened
rapidly to its upper and broken extremity, and acquired the same remarkable breadth
which is so conspicuous a feature of the scapula of Diplopias. The spine of the scapula
was also very oblique, inclining outwards as in the scapula figured in Plate XXXV.
The fossa glenoidea had precisely the same exceedingly circular outline as is seen in
the figured scapula; the coracoid process did not project much, and was recurved in
the same characteristic manner. On the outer margin of the neck, however, where
I found a deep fossa in Diplopus^ the scapula from Puy presented only a flattening.
In general the resemblance was as great as could be between two animals belonging to
the same family but to different genera.

The Humerus. — I have been able to study many specimens of humeri belonging to
Hyopotamus from the Isle of Wight, as well as from Puy, but unfortunately not a single
complete one. As is generally the case with fossil humeri, their upper or proximal head,
being very spongy, is destroyed during the process of fossilization, while the distal extre-
mity is well preserved. Compared with a humerus from Hordwell, the humeri from Puy
and Hempstead proved entirely similar to it ; and as the Hordwell specimen, belonging
to Diplopus, was the best preserved, I have figured it on Plate XXXVI. fig. 4, and
my description of this humerus will apply equally well to both genera. As mentioned
before, the proximal heads were broken in all specimens ; but seeing the Anoplotheroid
affinities presented by the distal extremity, we may presume that the proximal head also
resembled rather the Anoplotherium than the Suidse. In the first genus, as far as can be
judged by a crushed Anoplotherium humerus in the British Museum, the great tuberosity
did not overarch the bicipital groove so much as it does in Pigs and the Hippopotamus.
We may, to a certain extent, infer the lesser overarching of the great tuberosity by
marking the course followed by the crista anterior descending from this tuberosity,
and which in Hyopotamus runs in the middle of the anterior surface of the humerus
(fig. 4), and not so much on the inner side of it as in Suina. The deltoid ridge meets the
crista anterior a little higher up than in Anoplotherium, nearly as in the Hog ; and at the
point of their meeting we see a conspicuous rugose flat surface for muscular attachment.
The shaft of the humerus belonging to the Diplopus is very stout, thicker transversely
than the humerus of a Eeindeer, and much stouter in antero-posterior depth. The
transverse section is not so regularly oval, but much more triangular than in Pigs or
Ruminants, with the apex of the triangle turned forwards.

The inferior extremity of the humerus is very unlike that of any existing Ungulate,
and presents a good intermediate form between the humerus of Anoplotherium and that

OSTEOLOGY OF THE HYOPOTAMID.E. 60

of the Suina. As this inferior extremity of the humerus seems to me a very important
part, especially as showing the modifications which a bone may undergo in adaptation
to different conditions of life and organization, I must describe it at length.

Looking at the distal extremity of the humerus in the two series of fossil Ungulata,
we shall see that its shape is at first exceedingly and typically different in Paridigitata
and Imparidigitata. In the oldest Imparidigitata, as Palmotheriwn^ and even in the
living Rhinoceros, the distal extremity of the humerus is quite hourglass-shaped — that
is, it looks as if two truncated cones were joined together by their apices. At the point
of meeting we have a middle groove, from which the two horizontal cones thicken
gradually in both directions, inwards and outwards. As we have every reason to
suppose that Anoplotherium is the descendant of a very old type of Paridigitata, we may
look to its humerus as giving us the typical form of the distal extremity of this bone
in old Paridigitata. Now the distal extremity of the humerus of Anoplothermm is
totally different from what we have seen in Pakeotherzum ; the difference may be best
imagined if we say that in lieu of the middle groove, where the two cones meet, we
find in Anoplotlierium a round bulging, which goes all round the distal extremity of the
humerus (see Rlainville, Ost. Anopl. pi. iii.). This middle bulging is very characteristic of
all mammals in which the humerus is very movable upon the two bones of the antibra-
chium — so in Man, in most Carnivora, and in those Eodents which use their fore paws as.
hands. With the reduced mobility of the humerus upon the antibrachium, we remark
a concomitant change in its distal extremity ; the middle bulging recedes gradually to
the outer half of this extremity and becomes much sharper, till at last, in animals with
greatly reduced limbs whose humeri are fitted only for a simple sliding movement in
one vertical plane (as in our modern Ruminants, and in some Eodents, as Hares), this
round bulging of the Anoplothermm is reduced to a sharp ridge, which enters deeply
into a corresponding groove on the proximal extremity of the radius (which must
necessarily be adapted to all the modifications of the humerus). It is exceedingly
interesting to follow this gradual modification, step by step, through all the inter-
mediate stages presented by the Choerothermm of Lartet, the Hyomoschus crassus, to our
modern Ruminants, whose humeri show, instead of the Anoplotheroid bulging, an
exceedingly sharp ridge fitting closely into a corresponding groove of the radius, and
preventing any other movements save those in one vertical plane. It is interesting to
notice that we see the same change going on in the Imparidigitata, though their
starting-point, from the simple hourglass-shaped form of the humerus of the Palceo-
theridw, is so different. In this series a rising appears gradually on the outer half-cone :
in Palceotherium medium ; this rising goes on increasing with the gradual reduction
of the free movements of the fore limb in AncMtlierium and ffvpjparion, till it reaches
the state in which we now find it in the Horse ; and in this last the distal extremity of
the humerus is nearly like that presented by Ruminants; but as I have discussed this
case in my memoir on AncMtlierium *, I will not return to it in this place.

* Mem. Academie de St. Petersbourg, 1873.

BE. W. KOAYALEYSKY ON THE

In the humerus of Hyopotamus and Diplopus the middle Anoplotheroid bulging
(Plate XXXVI. fig. 4, a) is much lowered, showing a step toward the condition seen in
Pigs, in which it is raised into a very slight eminence. The inner margin of the humeral
extremity (fig. 4, h) is produced a little downwards, but much less so than in Anoplo-
therium. The supinator ridge is not prominent, though its rugose surface testifies to
the attachment of strong muscles. There is a very large intercondyloid perforation (e),
into which the olecranon entered deeply, as in modern wild hogs. The breadth of the
articular surface is very great comparatively to the transverse diameter of the whole
distal extremity of the humerus. The humerus figured in Plate XXXVI. fig. 4 belonged
to the didactyle form, Dijplopus Aymardi ; it is from Hordwell : besides this humerus I
had many specimens from Puy and Hempstead ; and as the two genera are confined to
their respective localities, the Diplopus to Hordwell and the Ilyopotamus to Puy and
Hempstead, there is no danger of intermixing their fossil remains. The humeri of
different size, which belonged to the tetradactyle Ilyopotamus, are entirely similar to
the one figured ; the only difference which may be noticed consists in the fact that the
distal articular surface is relatively larger in the Dvplojpus than in Eyopotamus, as seen
by the fact that in the three humeri of nearly equal size measured by me, the breadth of
the articular surface is 42 millims. in Diplopus, and only 32 and 35 in two humeri of
Ilyopotamus.

Dimensions of the Humerus.

Biplopus
Aymardi.
(Fig. 4.)

Hyopotamus.

Hordwell.

Hempstead.

Puy.

Largest transverse diameter of the distal extremity. ....
Transverse breadth of the articular surface

• • •

54i

42~

55
31

27|

28
22g
20
20

Yertical height, articular surface, internal border (b) . . .

Yertical height, articular surface, external border

Yertical height at the middle groove (a)

Transverse diameter of the shaft (inf. -J-)

Anteroposterior diameter

The Ulna. — The antibrachium of Hyopotamidce consisted of two completely separate
bones, and there is no trace of their having been immovably connected together, as in
nearly all living Ungulates. I had several specimens of this bone from Puy, and from the
English localities of Hempstead and Hordwell ; all, however, were more or less broken,
save a splendid specimen from Hordwell in the British Museum. This right ulna (Plate
XXXVI. fig. 1) fits the humerus figured in the same Plate (fig. 4) as exactly as if it came
from the same individual. I will give the description of this complete specimen, and
state the differences it presents from other ulnee from Puy ; the complete specimen figured
in Plate XXXVI. belonged to the didactyle Diplopus. The shape of this bone in our
didactyle genus is very striking from its extreme flatness and breadth ; it is much arched
forwards, and this curvature reminds one of the ulna of Suinae. At the upper part we see a

OSTEOLOGY OP THE HYOPOTAMIMS. 35

very large and broad olecranon, which differs slightly from living Euminants and Pigs by
its greater transverse breadth, and by the more projecting hook of the fore part, which
entered deeply into the intercondyloid perforation (fig. 4, c) of the humerus. The fossa
sigmoidea is uninterruptedly united on the inner side to the radial facet, while its outer
part presents a deep Schancrure. The outer radial facet of the ulna is not united to
the fossa sigmoidea in the Dvplopus, though it is so in the smaller Hyojpotamus from
the Isle of Wight and from Puy (fig. 2, or). Both articular facets for the radius are
nearly in one plane, whilst in Euminants the outer facet projects much more forwards
than the inner. The section of the ulna of Diplojpus gives us a figure like a flat or low
triangle, not quite regular on its external border. The posterior surface (Plate XXXVI.
fig. 1 ",_£>) of the ulna is the large basis of the triangle ; the obtuse apex (a) presents a
rugose ridge, seen in fig. 1/, running on the fore part of the ulna, and by which it is pressed
against the radius; the sides of the ulna are inclined planes, uniting the extremities of
the base to the apex. Just below the second third of its length the three planes of the
triangular ulna begin to narrow whilst descending to the distal extremity, which retains
the same triangular outline. The inferior extremity is cut nearly at right angles by the
facet for the outer bone of the carpus.

If we compare smaller ulnae from Puy and the Isle of Wight we shall see some dif-
ferences not only in the shape of the sigmoid fossa, but likewise in the transverse section
of the ulna. In some of the ulnae, like the one figured from Puy (fig. 2), the lower and
fore part of the fossa sigmoidea is much produced forwards, forming a sort of prominent
bridge between the inner and outer radial facets of the ulna ; the production of this
connecting bndge creates a deep fossa in the middle of the anterior surface of the ulna.
Having examined this part of the ulna in a large number of recent Suina, I found it
very variable, and therefore cannot lay too great a stress upon it in the Hyopotamidce.
But besides this, the ulnae from Hempstead and Puy, belonging undoubtedly to Ilyo-
potamus, have a very different horizontal section, and they are by no means so much
flattened as the ulna of the Diplopus ; this difference is most clearly seen by comparing
the horizontal sections of both ulnae taken about the middle of the bone. The species
from Puy haive a much deeper and sharper posterior edge (fig. 2', jp), which is nearly
absent in the Diplopus, Unfortunately I have not found a single entire ulna of the
Hyopotamus, though I have seen a large number of broken specimens from Puy and the
Isle of Wight. Comparing these broken ulnae with the one figured, I found some
notable differences in details, which will be clear to the reader by comparing the figure
of a part of the ulna from Puy and its transverse section (figs. 2 & 2') with that of the
ulna from Hordwell (figs, 1 & F). The differences extended further than the shape of
the sigmoid cavity, as the following Table of measurements will show.

MDCCCLXXIII. g

DE. W. KOWALEVSKY ON THE

Dimensions of the Ulna.

•

Jtiplopus

(%• h
Plate XXXVI.).

. Hyo^ot

amus.

Puy (Plate XXXVI. fig. 2).

Hempstead.

Total length along the curvature

270

■

Ju Uto Lt?X L\JL vXXL'X UL . < T • • » « •• • » • • »•♦»■•• • » •»■••.

254

Greatest breadth at the radial facets . ,

Breadth at the middle . . . ♦ .

* *

•" •

Antero -posterior diameter (at radial facets) . .

/jsft

Antero-posterior diameter (middle) ........

» »

• *

Transverse breadth, inferior extremity ......

Antero-posterior diameter, inferior extremity

11*

Antero-posterior diameter from the point 1
of olecranon to the posterior surface .... j

* •

• *

* •

The distal end of the ulna of Diplopus is triangular, with a slightly excavated tri-
angular facet for the pyramidale of the carpus, I have not the pisiform bone ; but as
there is no facet on the posterior side of the distal end of the ulna, this bone probably
did not articulate with the ulna, but exclusively with the pyramidale.

The great difference in the horizontal sections of the ulnee of both genera (fig. 1" and
fig. 2', Plate XXXVI.) is produced by the great flattening of the ulna of the didactyle
Diplopus. The section of the ulna of a hog taken in the middle will give nearly the
same figure as the section of the Hyojpotarnus ulna (fig. 2 1 ). The same letters mark the
corresponding parts in both sections. As an instance of a greatly compressed ulna, I
may adduce the ulna of Hyomoschus; it is so compressed laterally that its antero-
posterior depth is ten times as large as the transverse breadth ; and the whole bone, from
its radial articulation downwards, looks like a knife-blade with its sharp edge turned
forwards towards the radius.

The Badius (Plate XXXVI. fig. 3). — I was not fortunate with this bone, as I have not
a single complete specimen ; even fragments are rare., I found, however, in the collection
of the British Museum an upper half and a distal extremity of a radius from Hempstead
which belonged undoubtedly to Hyojootamus. I possess also the same parts from Puy,
and they entirely agree with the English specimens.

The proximal extremity of the radius is always shaped so as to fit the distal end of
the humerus, and their variations are always correlative. For this reason the radius of
Anofllotherium, constructed to fit the very peculiar distal end of its humerus, is made on a
pattern exceedingly different from that of all other Ungulata, and shows a striking likeness
to the proximal extremity of the radius of a carnivore, especially the Dog; but in Jlyo-
jpotamus we find a more Ungulate-like radius. Instead of the great oblique middle
fossa of the Anoplotherium, we see in Hyopotamus a shallow and broad groove (fig* 3, ol)]
made to fit the middle bulging (fig. 4, a) of the distal extremity of the humerus. The
inner part of the proximal surface is a plane (Plate XXXVI, fig. 3, b ! ) less inclined than
in Anoplotherium (Blainv. Ost. Anojpl. pi. iii.), but much more so than in Pigs, owing to
the greater downward production of the inner condyle of the humerus (Plate XXXVI.

OSTEOLO&T OF THE HldPOTAMID^. 37

fig. 4, b). The two posterior small fossae for the radial facets of the ulna are not very
deep; they are separated by a posterior bony projection, which entered, or is thrust
under, the connecting bridge of fig. 2, cb. We have no direct means of ascertaining the
curvature of the radius, but it is given to us by the shape of the ulna (fig. 1), so that, at
least in Diplopus^ the radius must have been considerably arched forwards.

The distal extremity (fig. 3 ; ) of the radius is very unlike that of any existing Pari-
digitate. In these we see generally on the distal end of the radius two longitudinal
excavated facets, separated by an- oblique prominent ridge running in the interspace
between the scaphoid and the lunare (see Blainv. Ost. Anopl. pi. iii.) ; we have no trace
of this ridge in Hyojpotamus. The inner half of the distal end (fig. 3', i) is occupied by
the oblique convex facet for the seaphoideum; the outer half presents a concave
anteriorly broad facet for the lunare; this bone, having a very oblique position in the
carpus, encroaches by its posterior narrow prolongation upon the scaphoidal half of the
distal extremity of the radius. : Both facets are separated anteriorly by a deep groove,
seen opposite the number 3' of the figure. The difference from the radii of all other
Ungulata is considerable. To the external straight truncated surface (fig. 3', ex) the
distal extremity of the ulna was articulated ; but I have had no specimen of Hyopotamm
showing this distal end of the ulna.

The radius of Hy V oU m «s was articulated (as in all Ungulates having a completely
developed ulna) only with the two inner bones of the carpus, the scaphoid and the lunare,
while the outer bone, the pyramidale, is taken up entirely by the distal extremity of
the ulna. AH existing Suidse show us the same relation ; but in Dicotyles, whose skeleton
is a little more reduced than that of the typical Pigs, we see that the distal extremity of
the radius grows broader in consequence of the reduction of the ulna ; and, besides its
two typical carpal bones, makes an encroachment upon the outer one, the pyramidale.
With the still greater reduction of the ulna in most Ruminantia, the radius goes on
increasing, and not merely touches the pyramidale, as in Dicotyles, but takes the whole
half of its upper surface for its own support, while the reduced ulna is pushed back to
the posterior half of the pyramidale. In consequence of this changed relation between
the carpal bones and the radius, the distal extremity of this bone in Ruminants acquires
&t its outer border an additional facet for the pyramidale, of which not a trace is to be
seen in Hyopotamus. We may mention as another peculiarity of the distal end of the
radius of Hyopotamus^ the entire absence of any styloid process or prolongation of the
inner border of the distal end of the radius, which is well seen in other Paridigitates, as
Anojplotherium and Hippopotamus. Specimens of the distal and proximal extremity of
the radius which I possess from Puy agree in every particular with the one described
from Hempstead.

ck2

DE. W. KOWALEVSKT ON THE

Dimensions of the Radius.

•■■•■-— , ■' ■—■■■■ ' ■ —

Hyopotamus.

Hempstead.

Puy.

Transverse breadth of the distal extremity. These ]
extremities seem to be epiphyses of radii of y

32
19

The Femur. — Of this bone I had a complete specimen from Puy (fig. 5, Plate XXXVI. ),
an upper half of a smaller species from Hempstead (fig. 6), and a lower half
(Plate XXXV. fig. 2) from the same locality, but belonging evidently to a larger indi-
vidual or species than the upper half. I had no femur from Hordwell, and consequently
this bone is unknown in Diplopus.

Confining our comparison only to the Paridigitate series of Ungulata, of which Hyo-
potamus is one of the old representatives, we find in our fossil femur characters that are
very common to all the members of this division. In comparing the ruminant and non-
ruminaiit Paridigitata, we find that the femur in the Suina has a more complete
spherical head supported on a pretty distinct neck, the great trochanter rising very
slightly or not at all above the level of the femoral head, and a small (or inner) tro-
chanter not very prominently developed, and often consisting merely of a rugose thick-
ening at the antero-posterior edge of the superior half of the femur. The general shape
of the bone is very round in Dicotyles, more flattened in other pigs, and slightly arched
from behind forwards. Both edges of the anterior part of the distal articular surface
(rotular surface) are alike, while the internal condyle is thicker than the external ; the
transverse breadth of the trochlear surface for the patella is proportionately broader in
comparison with the entire thickness of the distal end than in Euminantia (Dicotyles
forms an exception to this rule, and the rotular breadth of its femur presents the same
relation to the whole thickness of the distal end as in Euminantia). Now the femur of
Hyopotamus shows characters that are common to both divisions of Paridigitata. The
head of this femur is supported on a neck (fig. 6) even more distinct than in the Suina,
and approaching that of Hippopotamus. The great straightness of the whole femur
reminds us also of this last genus/ The spherical head is provided with a deep pit for
the round ligament, which is absent in Hippopotamus. This round articular head is
connected with the prominent small (inner) trochanter by a very sharp ridge, as in some
Suina, only the ridge is higher and sharper (fig. 6'). The bridge of bone (fig. 6, a) con-
necting the head *of the femur with the great trochanter is much contracted in the
middle, and lowered in such a way that both articular head and great trochanter rise
considerably above the level of the connecting bridge — a character very general among
the Paridigitata, with the exception of the Camelidw, in which the superior end of the
femur is shaped on a plan entirely different from other Euminants, presenting a great
resemblance to the Imparidigitata (Bhinoceros) in the breadth of the connecting bridge

oxJiUJjUljrx Ur XxLsh JAXUxUlAMlJUzlif*

and the shape and position of the great trochanter. The shape of the femur in
Hyopotamus (fig. 5) is remarkably uniform in its thickness, is exceedingly straight, and
presents on its posterior surface, above the internal condyle, no deep pit for the plantaris
muscle, but a simple rugosity, as in hogs. Some comparative anatomists seem to attach
a great importance to this fossa, though I cannot do so, seeing that this fossa
is developed or absent in animals belonging to both series, Paridigitata and Impari-
digitata. For instance, it is very large in the Horse, but absent in the Rhinoceros ;
developed in most Ruminants and Hippopotamus, but absent in Camelidw and Suidce^
where we find, in place of the pit, a rugose surface for the attachment of the same
muscle.

The fore part of the distal extremity (Plate XXXV. fig. 2), as in most living Ungulates
of both series, with the exception of the Suinse, Hyomoschus, and some Ruminants, has
a more developed internal rotular edge, though the difference in thickness of the internal
and the external edge is not carried to such a degree as in Bovidae and Horses, but is
comparatively slight. The internal condyle is also a little thicker than the external.

The lower extremity (Plate XXXV. fig. 2) of the femur from Hempstead may belong
to Hyopotamus bovinus by size : this last specimen was Idndly lent me by the Museum
of Practical Geology in Jermyn Street ; it is figured of the natural size.

Dimensions of the Femur of Hyopotamus.

Puy.

Hempstead.

Hempstead (Plate XXXV. fig. 2).

Transverse breadth of the"
proximal extremity be-
tween the articular head j
and the great trochanter . . J

283
62

47'

Transverse breadth of the upper broken end .. . 36

Antero-posterior depth 25

Transverse breadth of the distal end ........ 6 1

Breadth of the rotular surface 26

The Tibia (Plate XXXVI. fig. 7). — I had for the study of this bone two nearly entire
specimens from Hordwell, belonging to the large didactyle form Diplopus, and some
incomplete upper and lower ends from Puy ; these last undoubtedly from the tetra-
dactyle Hyopotamus, as no Diplopus is found at Puy. The tibia of the Diplopus from
Hordwell approaches very nearly iii length to that of a Keindeer, showing that the
two-toed Diplopus must have been very high on the hind legs, and the long metatarsals
confirm this view. The general shape of this tibia is triangular from the upper
part down to the distal extremity, and not so much rounded in the lower half as the
tibia of the Suinse. The crista anterior is not very high, with only a slight patellar

depression on its fore part, much shallower than in Pigs. The shaft of the whole
bone is very straight ; the outer edge, facing the fibula, is exceedingly sharp, while the
inner is more rounded. The distal end (fig, 7') shows a very rectangular outline, with
two straight deep grooves for the upper pulley of the astragalus. The inner distal process
of the tibia, or the so-called inner malleolus, which holds the astragalus from the inner
side, is much longer than in Suina, approaching more to the size it exhibits in most
Ruminants, with the exception of Tragulina, as these last have an exceedingly short

DR. W. KOWALEYSKY ON THE

inner malleolus. The outer distal edge is truncated somewhat obliquely, and has a;
groove in the middle for a corresponding convex ridge on the inner surface of the
fibula, which fits the tibia in the same way as in Pigs, and being produced distally
together with the internal malleolus, firmly clasps the astragalus (figs. 7 & 7', Plate
XXXVI.). The posterior surface of the upper half of the tibia is very flat, showing
some oblique, rough bony ridges for muscular attachment.

The upper or proximal articular surface for the femur presented more resemblances
to Suinse than with Euminantia, especially in the deep outer groove for the muscular
tendon, though the patellar fossa on the fore part of the crista anterior is shallower
than in Sus.

In the broken upper and lower halves of the tibiae of Hyopotamus I find no difference
from Diplopus, save that of size.

Dimensions of the Tibia.

. Diphpus
(fig. 7,

Plate XXXVI.).

Hyopotamus;

Hordwell.

Puy.

Hempstead.

Length.

280
68
43|

240

55
31£

Transverse breadth, proximal extremity . . . .
Transverse breadth, distal end

The Fibula. — Before describing this bone in the Hyopotamidce, we must call to mind the
principal differences it shows in the two series of Paridigitata and Imparidigitata. The
proximal end is wanting in all our specimens ; but this is not of any consequence, as all
the chief characters are presented especially by the distal extremity. In all Impari-
digitata which possess a complete fibula, this bone is applied to the outer side of the
tibia, forming at the distal extremity the malleolus externus (Plate XXXVI. fig. 7) ;
this distal extremity is truncated obliquely in such a way that the inner surface of it
articulates only with the astragalus, forming its outer boundary; the fibula can never
touch the calcaneum*, as this bone has no special surface developed on its upper lateral
part to receive the distal end of the fibula. On the contrary, in all Paridigitata the
fibula being applied in the same way to the tibia, presents at its distal end two articular
facets for the articulation with two tarsal bones. The inner surface of the fibula
(Plate XXXVI. figs. 7 & 7', and Plate XXXV. fig. 3, nat. size) is pressed against the astra-
galus ; but its inferior extremity does not thin out, as in Imparidigitata, but is truncated
at right angles, and provided with a special articular facet for an articulation with the
outer wall of the calcaneum. We shall see hereafter that this outer wall exists
perhaps also in the Imparidigitata, but that it is thrust under the astragalus ; and this
seems to be the cause why the fibula of the animals belonging to this series cannot
articulate with the calcaneum.

* Macrauchenia seems to present the single exception to this rule. The calcaneum of MacraucTienia is not

known.

OSTEOLOGY OF THE HYOBOTAMID.2B.

Ihe fibula of Dipiopus (Plate XXXVI. fig. 7, and Plate XXXV. fig. 3) is shaped
entirely in conformability with the Paridigitate type; it is a considerably reduced bone,
with a very thin shaft; the distal extremity, however, is widened considerably in the
antero-posterior direction. The inner surface of this broad (it would be more correct
to say deep, as the fibula is widened from before backwards) distal extremity (Plate
XXXV. fig. 3) is shaped on the same pattern as in the Suidse, presenting a more uniform
arched and slightly raised platform pressed against the astragalus, and not a deep semi-
circular notch as in Euminantia. (In these the shaft of the fibula is generally mostly
reduced to a mere tendon, but its lower extremity remains under the name of the so-
called " osselet peroneen") The articular end for the calcaneum (fig. 7') presents a very
deep and narrow facet, which occupies the whole antero-posterior extent of the distal
extremity of the fibula. The anterior part of this calcaneal surface is slightly convex,
the posterior concave.

I have found at Puy many distal extremities of the fibula, which belonged undoubtedly
to Hyopotamus ; and even this bone bore a great resemblance , to the same part of
Dipiopus. This is another fact testifying to the very close relationship of both genera,
notwithstanding the difference in the number of toes. A resemblance which holds
good even in such slight details is a conclusive proof that both genera — the Dipiopus
from Hordwell and the Hyopotamus from Puy (and Hempstead) — belonged to the same
family, and that one may be considered the reduced form of the other, notwithstanding
the seemingly adverse fact that the reduced form is met with in the Upper Eocene, and
the more complete in the Lower Miocene. This last certainly had Eocene ancestors,
which may be considered to have given rise to the reduced form, while they continued
to live themselves until the Miocene period.

Dimensions of Fibula, distal end.

Dipiopus

(Plate XXXV. fig. 3,

nat. size).

Antero-posterior diameter of the upper broken end. .
Thickness of the same ............... ...'...,,,

Antero-posterior diameter of the distal end

8
4

General considerations on the carpal and tarsal, metacarpal and metatarsal bones in

Hyopotamus, Dipiopus, and other Paridigitata,

If, generally speaking, the long bones of the limbs in Ungulata often present but
few characters decisive enough to tell us at once the natural series to which an Ungulate
mammal belongs, the converse is the case with the smaller bones of the extremities, which
have therefore a great systematic importance. We may very often know most of the long;
bones of the skeleton, the scapula, the humerus, the antibrachium, the tibia and fibula
of a fossil Ungulate, without being able to determine quite certainly the natural series
to which it belonged ; nay, even more, we may discover the skull and the complete

42 DE. W. KOWALEVSKY ON THE

dentition without being made much wiser by it. The history of palaeontology swarms
with such examples ; but the discovery of a single carpal or tarsal bone very often
clears the whole question, by showing in the most unmistakable manner the true
affinities of a fossil form. Seeing this great importance of the carpal and tarsal,
metacarpal and metatarsal bones, I feel obliged to enter into more minute details than
I have done in the case of the long bones of the limbs ; and I shall try to show how
constant and important are the characters which we may derive from the study of these
small bones, and what excellent data they may furnish towards a complete understanding
of the development in time of the large and extensive group of modern Ungulata, As
in one of my former memoirs* I tried to follow this course in reference to the Impari-
digitata, I will confine myself in this paper only to the Paridigitate series. As the chief
differences for subgeneric or specific division of the Hyopotamidw are furnished by the
bones of the feet, I shall describe these first, leaving the skull and the dental characters
to be treated afterwards.

Although I found at the British Museum and in the private collection of M. Aymaed
at Puy very extensive materials for the restoration of the fore and hind feet, still it
is to be regretted that as yet we have never found a complete fore or hind foot, in its
natural connexion, belonging to the same individual ; still less was there a chance
of finding the bones associated in an undoubted manner with a certain set of teeth : all
bones of the Hyopotamus occur very much scattered, and no complete skeleton belong-
ing to one individual has ever been found, at least to my knowledge. The happy cir-
cumstances which enabled Cuviee to refer, without any doubt, certain sets of bones to
certain skulls (as whole skeletons were sometimes found in the gypsum of Montmartre)
did not favour my research. But if the actual connexion of specific bones and teeth
should still remain not entirely cleared up, the general osteology of the genus will not
in the least suffer by it ; and with the materials I had at my disposal, I am conscious of
being able to reconstruct both generic forms in a very satisfactory manner, though
even for the carpus I had no complete set of bones belonging to one individual. But in
certain series of animals the peculiar form and general relations of these bones are so
constant, that we may expect variations only in the smallest particulars ; so that the
general structure of the feet and the mutual relation of the carpal and tarsal bones
between themselves, and to the bones of the metacarpus and metatarsus, may be con-
sidered undoubtedly settled for the Hyojpotamidce. To make my description more
clear to the reader, I will take care to make comparisons with animals accessible to every
naturalist; and I think that for the full comprehension of the relations of these
irregular bones drawings are wholly inadequate, and a direct comparison with the feet
of a pig and a ruminant should be resorted to.

Notwithstanding the great external diversity of the recent Ungulata, taking both
series of Paridigitata and Imparadigitata, where we meet with animals so different in
their aspect, habits, and size, running through all the intermediate stages from a rabbit

* " Sur rAncMtherium," Mem. Acad. St. Petersburg, 1878,

OSTEOLOGY OP THE HYOPOTAMHX3E. 43

{Hyrax) to the largest Rhinoceros, Hippopotamus, or Giraffe, their osseous structure is
exceedingly uniform and simple, presenting only two chief types, which are completely
defined by the terms Paridigitata and Imparidigitata. Beginning with the oldest
Eocene deposits, these two types are entirely separated by the structure of their skeleton,
and we have not a single living or fossil form that could truly be considered a link
between these two typically different series. I have little doubt that all Ungulates
must have sprung from some common form; but if so, their division was effected in
very ancient times, perhaps in the early Cretaceous period. I say early, because just at
the close of the Cretaceous period, in the lignites of Soissons and the calcaire grossier,
or in beds that are contemporaneous with them, we find large Mammalia of both
series with a very reduced skeleton. Considering the rate at which the reduction of
the limbs in Ungulates proceeded in the Eocene, Miocene, and Pliocene periods, we are
obliged to grant a long time for the branching off of the two series of Ungulates from
a common pentadactyle form, and the subsequent reduction of each branch to the three-
and two-toed forms, met with in ancient Eocene beds.

In the recent period, notwithstanding the great diversity and very wide distribution
of Ungulates, they present an extreme poverty of type. The Imparidigitate series is
strikingly poor, not only in generic but also in specific forms ; they go on evidently
declining from the Eocene ; and in the recent period we have only three different types
of these animals — the Equina, Bhinocerotina, and Tapirina; and even these three
present such a fundamental resemblance in the structure of their limbs and skeleton,
that no reasonable doubt can be entertained as to their descent from a common
progenitor. On the other hand, the Paridigitata, though presenting in the recent
period an exceedingly great diversity of size and habits and a great variety of specific
forms, outnumbering tenfold those of the Imparidigitata, present also only two distinct
types, Ruminantia and Suina; and, even in these, the fundamental structure of the
limbs and skeleton is so uniform that we may safely infer their descent from one
common form. In the present memoir I have taken up only the second series, the
Paridigitata ; and I shall try to show, as clearly as possible, the common bond that holds
them all together. I shall point out that the uniformity of this fundamental type adapted
to different conditions of life is so great that, notwithstanding extreme diversity of
size, difference of habits, aquatic or terrestrial life, we may trace through the whole
complexity of these diversified forms not only the number and shape of their carpal
and tarsal bones, but even each separate facet of these bones, and point it out as clearly
in the reduced limb of a land Antelope as it is displayed in the complete unreduced
limb of the aquatic Hippopotamus. Further, I shall try to show that this uniformity
of type holds good, not only of the living, but also of the extinct forms of Paridigitata.

If we cast a glance on the first section of the fore foot, or the carpus, of any
Paridigitate whose skeleton has not suffered too considerable a reduction (as for
instance a Pig or Hippopotamus, Plate XXXVII. fig. 1), we shall find that it consists
of two rows, containing seven small irregular bones, with an additional one (the

MDCCCLXXIII. H

44 DE. W. KOWALEVSKT ON THE

pisiform) on the outer end of the upper row. We find in the upper row, beginning
from the inner side, the scaphoid, lunar, pyramidal, and pisiform ; in the lower, trape-
zium, trapezoid, magnum, unciform. In all extinct Paridigitata, even those with a
very reduced skeleton, we find the full number of these carpal bones (so in Anoplothe-
rium, Xijphodon, Hyopotamus, Plate XXXVII. figs. 2, 3, 20) ; and, only in consequence
of the still greater reduction of the limbs, in some recent Ungulata one of the carpal
bones (the trapezium) seems to be entirely lost, while another (the trapezoid) becomes
confluent with the magnum ; that the trapezoid is not lost but is confluent with the
os magnum, is shown! by the two different points of ossification in the cartilage of the
os magnum in young sheep. As we shall hereafter see, the number and shape of the
carpal and tarsal bones of all animals belonging to the Paridigitate series present an
extreme uniformity, individually as well as in their relation to one another and to
the metacarpal and metatarsal bones ; and this similarity is indeed so great that we
cannot explain it in any other way than by community of descent. The extreme con
stancy in the relations of these bones in all Paridigitata being ascertained, the problem
which is unavoidably presented to the mind of the observer may be stated thus : — Very
irregular small bones, intended to constitute a movable articulation between the long
bones of the extremity and the metatarsals and metacarpals, arranged themselves in a
certain way in reference to one another and to these metacarpals and metatarsals;
this arrangement remains the same in all Paridigitata, recent as well as fossil,
notwithstanding the greatest diversity of form, size, and habits of life ; and if some
slight change is to be seen, it is due clearly to the overdevelopment of certain digits
and consequent reduction of others ; but in all cases the reason of change is at once
apparent : how can such similarity in animals so entirely different be explained 1 To
all naturalists who accept the gradual descent and differentiation of all Paridigitata
from one common form, the fact must appear as a perfectly reasonable and intelligible
one. If the immediate progenitor of the Paridigitates presented the given arrangement
of the carpal and tarsal bones, then, at the gradual differentiation of this type, every
small change in one bone called forth a corresponding change in all its neighbours. And
as the link which connects all the forms together was never destroyed, and the changes
were slowly going on, we meet now, in the extremely differentiated descendants, a unity
of organization which was inevitable if all these forms descended from one common
progenitor. But if, leaving the point of view of evolutionists, we look at the matter on
the special-creation principle, this similarity of structure in animals so widely different
is really an awkward fact. To the supporters of special creation the question presents
itself in its simplest form thus : — We have now on the earth a large assemblage of
Paridigitate mammals, presenting widely diversified generic and specific forms, fitted for
the most different conditions of life, some leading an amphibious existence, sharing the
large streams with Crocodiles, while others inhabit inaccessible rocks or burning sand
plains, some heavy and sluggish, others light and swift, &c. ; and yet the creative force,
in calling separately into existence these diversified forms, made them all on one plan,

OSTEOLOGY OF THE HYOPOTAMID^E. 45

and this to such an extent that even the seven bones of the carpus and tarsus, notwith-
standing their irregular shape, were always arranged in the same way, so that a certain
facet of one bone always touched a certain particular facet of another, and never other-
wise. That this could really occur in every separate case of creation, is almost as pro-
bable as that seven dice thrown out of a dice-box should give us the same number
of points, similarly arranged in a hundred successive throws. Notwithstanding the
thousands of different relations which might exist between such seven multangular bones,
we get always only one; and in the whole range of living and extinct animals we see no
exception to the common rule of typical arrangement of the carpal and tarsal bones.
The point at issue is, can this uniformity be accounted for by the principle of special
creation, or by the theory of descent and modification 1 No naturalist can in our time
hesitate between the two ; and while all the adduced facts are wholly inexplicable by
the first theory, they seem most natural in the light of the second. We may still not
be fully informed as to all the true causes which induced the variation and consequent
differentiation of animal types ; but the principle of descent must be conceded as the only
one by which all future researches into the structure of the extinct world must be
guided.

I have mentioned chiefly the carpal bones ; but the study of the tarsus leads to pre-
cisely the same result, and the likeness of the tarsal bones in all Paridigitata is perhaps
even more striking than that of the carpals. All Paridigitata have a calcaneum with a
special facet for the fibula, an astragalus with a double pulley, a cuboid supporting the
fourth and fifth digits, and a navicular, with three cuneiforms, for the support of the
third and second digits ; the first digit being always lost, its tarsal bone is gone to give
support to the second toe, or, if this be lost, to the rudiment of it. We shall see these
relations by-and-by, when we come to the special description of the tarsal and meta-
tarsal bones.

The Carpus of Hyopotamus, or the Four-toed Form.

In describing the bones of the carpus of the Hyopotamus^ I will try to confine my
comparisons exclusively to the nearest living relatives of the extinct genus, as only
such likeness and difference between nearly related forms belonging to one natural
series can be of any immediate use for our purpose. Resemblances to the bones of
animals belonging to other series are mostly only analogies, not homologies; and if
some similarities which we may find to the carpal bones of animals belonging to the
other natural series of Imparidigitates are of importance as testifying their common
descent from some ancient form, still we lack so completely any links between these two
series (which are entirely distinct from the oldest Eocene deposits) that it would be idle
to speculate about their relation on such trifling characters as these. I will therefore as
much as possible confine myself to the series of Paridigitata. The carpus of Hyopotamus,
like that of all Paridigitata, consisted of eight typical bones, four in each row; of these
eight bones I have only five, the trapezium, pisiform and magnum being absent from

46 DE. W. KOWALEVSKY ON THE

all the collections I had an opportunity of studying ; but their absence will not interfere
much with the complete restoration of the fore foot.

The scaphoid (Plate XXXVIII. fig. 5, s) is not very different from the corresponding
bone of the pig in its upper part ; but the general form tends more towards the Hippo-
potamus. The proximal surface reminds us of the same surface of the Anoplotherium
(Cuv. Oss. Foss. v. p. 217), and shows a very shallow platform, sloping radially ; in
correspondence with this, the inner part of the distal extremity of the radius (articulating
with the scaphoid) is also very flat. In our recent Suidse, this proximal facet of the
scaphoid has a great rising on the fore part and a deep excavation behind ; and, eorre*
spondingly, the extremity of the radius is more deeply excavated, and the interlocking
of the two is firmer.

The distal surface of the scaphoid in the Hyopotamus is very distinct from the same
surface in the pig, and reminds us more of what we see in Hippopotamus ; namely, it
is divided by an oblique ridge into two slightly concave facets, of which the radial or
posterior facet articulates with the trapezoid (Plate XXXVIII. fig. 5, t) 9 and the ulnar
or anterior reposes on the os magnum (Plate XXXVIII. fig. 5 and Plate XXXVII.
fig. 20), which being absent, its place is left blank in our figures.

The posterior extremity of the scaphoid is elongated into a thick recurved portion^
which bends inside the carpus. The inner or ulnar surface exhibits nothing particular ;
its upper margin is occupied by a long narrow facet for articulation with the lunare.
The outer free surface is uniformly rounded.

Dimensions.

Length antero-posteriorly ... 30 millims.
Height 16 „

The scaphoid (from Hempstead) described above belongs to the tetradactyle form of
the Hyopotamidce, or to the genus Hyopotamus ; but I have found in the collection of
the British Museum another scaphoid bone (from Hordwell), which I may with great
probability refer to the didactyle form called JDiplopus. This scaphoid is very different
from that just described ; and all the differences point to a reduction of the foot. Its
proximal surface does not slope outwards, but is perfectly horizontal, with a transverse
rising in the fore part, a hollow in the middle, and an elevated posterior border ; such
a surface of the scaphoid ensured a firmer interlocking with the radius. It reminds one
strikingly of the same surface in the Anchitherium ; and it is possible that the distal end
of the radius in the Diplopus was shaped on the Cameline type, which is very similar to
that of a Horse or Anchitherium.

The distal surface is not divided into two facets, but is uniform, like the distal surface
of the same bone of a pig ; and this is very intelligible, seeing that the trapezoid, having
no complete second toe to support (as in Hyopotamus), did not press on the distal
surface of the scaphoid, leaving there such an impression as it did in Hyopotamus.

The semilunar (Plate XXXVIII. fig. 5, e, and Plate XXXVII. I). — I had several

OSTEOLOGY OF THE HYOPQTAMID,£. 47

lunar bones of different sizes from English deposits as well as from Puy. Those from the
latter place agree entirely with smaller lunars from Hempstead ; and we may safely
infer that they belonged to the Hyopotamus.

In general shape the lunar has the same character as the same bone in other Pari-
digitata, especially the Suina— with this difference, that it is relatively much thinner
and higher in Hyopotarrms, as might be expected from the form of all the bones of
the skeleton, showing that the Hyopotamidse had a much more elegant and higher
skeleton than our recent Suina. The proximal or upper surface of the lunar, which
articulates with the outer half of the distal extremity of the radius, has a considerable
rising in the fore part and an excavation behind ; this upper surface, looking on it
from above, has a very oblique direction inwards; this is effected in such a way that
the radial upper margin of the lunar is pressed closely to the scaphoid, while there
is a great interval between its upper ulnar margin and the pyramidale : a glance at the
proximal surface of a pig's carpus will explain this disposition much better than long
descriptions.

The distal surface of the lunar is prolonged into a prominent beak inserted between
the os magnum and unciforme. This insertion of the lunar between the two principal
bones of the lower row is a feature common to all Paridigitata ; only in Hippopotamus
(Plate XXXVII. fig. 1, 1) the beak, owing to the squareness of all the bones, is much
blunted ; but it is very well seen in Anoplotherium, Xiphodon (Plate XXXVII. figs.
2 & 3, Z), the Suina and Ruminantia. The beak is limited on both sides by two oblique
facets, one radial for the os magnum, the other ulnar for the unciforme ; a ridge running
from the anterior point of the beak through the whole antero-posterior depth of the
bone separates the two facets from one another. On the anterior half of the distal
surface of the lunar, the ulnar facet for the unciforme is larger than the radial or
os-magnum facet ; but on the posterior half this relation of the two facets is inverse :
this disposition is obviously an adaptation for the better interlocking of the carpals.

The posterior termination of the lunar is very like that of the lunar of the Suidse :
the lateral facets for the two adjoining bones show nothing particular; only the upper
radial facet covers the whole upper and inner margin of the lunar, while the ulnar is
only developed on the fore part, the ulnar side not touching the pyramidal in its
middle and posterior part. I have, from Hempstead, a lunar bone much larger than
the one described and figured by me in the restoration of the carpus (Plate XXXVIII.
fig. 5) ; it may have belonged to the largest species, called Hyopotamus bovinus^ although
it looks too large for it ; there is no difference in shape whatever.

DE. W. KOWALEVSKT ON THE

Dimensions.

*»•> e • • • 0-

Height, anterior . .
Breadth, anterior . . .
Depth, antero-posterior. .

» *•• •. ». • « * * # « • »• « * « »

a • • • •

• »..**

Hyopotamus.

Hempstead.

19,22

Hempstead.

27£

28j

The unique specimen of the lunar I possess from Puy, although agreeing entiiely
with that from Hempstead, is somewhat too much crushed to admit of accurate mea-
surements.

The pyramidal (Plate XXXVIII. fig. 5, p) has the shape which is so characteristic of
it in all Paridigitata. We may remark, in reference to this bone, that it is the least
changeable of all the carpal bones, its shape being very similar, even in animals of
widely different families and even orders. For instance, it is very like in both
Paridigitata and Imparidigitata ; and a pyramidal of Ehinoceros is hardly to be dis-
tinguished from that of a Hippopotamus. We may suppose that, as this bone has a
very similar function to perform in all Ungulata, it did not change its shape, while the
other bones, more directly acted upon by the different condition of life, did.

The pyramidal of Hyopotamus presents therefore a certain likeness to that of an
Anoplotherium, and also to that of a common hog; the chief distinction lies in the fact
that all its edges are much sharper, and not so much blunted as in Suidee ; this applies
particularly to the ridge separating the facet of the ulna from the facet for the pisiforme
(Plate XXXVIII. fig. 5, p).

The distal surface, adapted to the outer proximal facet of the unciform, is shaped very
much as in all other Paridigitata.

I have this bone only from Puy.

Dimensions,
Height ........

Greatest antero-posterior depth .
Breadth

18*

The pisiform is wanting ; it is, however, not a very important bone, partaking more
of the character of a sesamoid than of a true carpal.

The second row of the carpus. — I could not find the trapezium in any collection, although
a facet on the radial side of the trapezoid and the second metacarpal clearly show that
it was present. As the first digit is always abortive in all living and fossil Ungulata,
the trapezium, which is the true carpal bone of this digit, has lost all its importance.
However, it is to be seen in Hippopotamus and Anoplotherium (Blaihv. Ost. Anopl.
pi. iii.), where it articulates with the posterior face of the trapezoid, and assists in
holding the second metacarpal in Hippopotamus, or the rudiment of the second in
Anoplotherium. It is present in the Suidse (Plate XXXVII. fig. 4, tz) ; and I found it

OSTEOLOGY OF THE HYOPOTAMHX&. 49

also in Phacochoerus and Babirussa ; it seems to hang only on the back part of the
trapezoid (#), without touching the second metacarpal. The trapezium is very small
in Bicotyles (Plate XXXVII. fig. 5, tz\ and is entirely lost in ruminants, not being even
present as a distinct point of ossification in the trapezoido-magnum cartilage. It is
worthy of notice that, though the metatarsus in all Ungulata is always more reduced
than the metacarpus, still the homologue of the trapezium (the first cuneiform) is often
found in the pes when no trace of its homologue exists in the manus. So, for instance,
in the Horse (coalesced with the second cuneiform) and most, if not all, Ruminants, in
which the first cuneiform is present, while the trapezium is lost.

It is difficult to say if it existed in the Diplopus, seeing the reduction of its meta-
carpals to two ; but as it exists in Anoplotherium there is no reason why it should be
lost in the two-toed Biplopus.

The trapezoid (Plate XXXVIII. fig. 5, t)>— I have this bone from Puy, belonging to
Hyopotamus or the four-toed form. In considering only the shape of this trapezoid,
without heeding the size, it is almost identical with the trapezoid of Hippopotamus;
only its upper or proximal surface is more convex. The distal surface is deeper than it
is broad, slightly concave, and fits exactly the proximal surface of the second metacarpal
(Plate XXXVIII. fig. 5, t). On its radial side is seen a well-developed articular facet
for the trapezium. If we compare this trapezoid with that of a pig (Plate XXXVIL
fig. 4, t\ we shall find a great difference, especially in the distal surface. In Hyopotamus,
as well as in Hippopotamus, the trapezoid is destined to support only its typical
metacarpal (the second), while in the Suidae, owing to the larger development of the
middle^ and consequent reduction of the lateral metacarpals, one half of the distal
surface of the trapezoid is taken by the third metacarpal, and only the remaining
half supports the second (Plate XXXVIL fig. 4, t). In consequence of this, the
distal surface of this bone in the true Suidae, instead of being flat, is spear-shaped, with
a sharp edge running through its whole antero-posterior depth. In Bicotyles (Plate
XXXVIL fig. 5, t), where the reduction of the lateral metacarpals has gone still further,
the distal surface of the trapezoid, being wholly taken by the third metacarpal, has
resumed its flat form, thus resembling more the trapezoid the Hyopotamus than that of
the pig. But notwithstanding this similarity, the part played by this bone in Bicotyles
and Hyopotamus is wholly different ; and while in the first it has no connexion with the
second metacarpal, in the last it is entirely connected with it, and gives no facet to the
third metacarpal.

Dimensions.

Height ......

• • JL \J a

Breadth, inf. ....

. 8

Antero-posterior depth

. . lo

Unfortunately the os maynum is wanting in all the collections I have visited ; but
the interval included between the surrounding bones gives an idea of its shape, which

50 DB. W. KOWALEVSKY ON THE

certainly must have approached that of Anoplotherium. I pass on to the most important
of the bones of the second row — the unciform.

Happily we are very well oif as to this bone, there being some excellently preserved
specimens from Puy. The shape of this bone and its connexion with the adjoining parts
present considerable interest.

The unciform of Hyopotamus (Plate XXXVIII. fig. 5 u, fig. 7) presents some like-
ness to the same bone of Anoplotherium and also of Sus ; but nevertheless its differences
from both, especially the last, are numerous. Comparing the front view of fig. 5 and
the view of the distal surface (fig. 7) with the corresponding aspects of the unciform
of Anoplotherium in Cuvier (plate 102, fig. ii. A 2 & 3), we shall see that they present
many features in common in both genera ; I shall add that in Cuvier's figure (A 3),
the letter h corresponds with our III., the letter h with our IV., and the letter i with
our V.*

Looking at the bone from above or in front (Plate XXXVIII. fig. 5, u% we see that
its radial facet articulates with the outer face of the lunar bone, while its ulnar facet
is occupied by the pyramidal. The facet for the pyramidal is larger than that for
the lunar, both in Hyopotamus and Anoplotherium (Plate XXXVIII. fig. 5, u, and
Plate XXXVII. fig. 2, u) 9 while it is just the reverse in Sus, where the lunar facet of
the unciform is much larger than the pyramidal facet. Both facets are divided by a
pretty prominent ridge. The whole bone is relatively higher than in Anoplotherium,
in correspondence with the lighter and more elegant stature of the Hyopotamus.

The distal articular surface is shown in Plate XXXVIII „ fig. 7, and may be compared
with the same aspect of the unciform of Anoplotherium (Cuv. pi. 102. fig. ii. A 3).
We find on our bone the same facets ; only their relative development is different.
III. is the radial inferior facet, giving articulation to the prolonged beak of the third
metacarpal (k in Cuvier) ; IV. is the middle, principal, or central facet, articulating with
the fourth metacarpal (h in Cuvier) ; outside from it is seen another smaller, facet (V.)
for the lateral or fifth digit (Cuv. i)f; and we see just the same in the Suidas, which
have a complete fifth digit, and in Anoplotherium, where there is only a rudiment of
it. In Hyopotamus and Hippopotamus (Plate XXXVII. fig. 1), however, owing to the
greater development of the fifth metacarpal, this facet is not pressed so much backwards
and upwards as in the pigs ; in the living Hippopotamus it is entirely on the same
level with the facet for the fourth metacarpal (Plate XXXVII. fig. 1, u).

* In Cuvier the bone is turned in an opposite direction, though it is from the same side, the left.

t Cuvier (Oss. Eoss. 4th ed. vol. v. p. 425), describing this facet i as supporting the rudiment of the fifth
toe, adds, " On n'en trouve ^analogue ni dans le bceuf, ni dans le cochon, ni dans le Tapir, &c." It is certainly
not found in Bos, as the Ox has no rudiment of the fifth toe ; but it is always to be found in the Suidse as well a£
in the Tapir, both these having a completely developed fifth toe, which is articulated to the homologue of this
facet. In all Mammalia, without exception, the unciform supports the fourth and fifth metacarpals ; therefore
in all mammals both these facets are homologous, whether they support completely developed toes or onlf
rudiments of such, as in Anoplotherium.

OSTEOLOGY OF THE HTOPOTAMID^E. 51

In examining the radial side of the unciform of Hyopotamus we shall see that the
upper lunar facet of this bone nearly meets the lower facet (Plate XXXVIII. fig. 7), the
two being separated merely by a ridge*, while in Anoplotherimn (Plate XXXVII. fig. 2, u,
and Cuv. plate 102. fig. ii. A 5 m) these two facets are separated by a vertical facet,
which articulates with the neighbouring os magnum. In the Suidw this facet is exceed-
ingly developed, and the homologue of the lower facet in. (Plate XXXVII. fig. 7) is
confluent with it in such a way that both together (k-\-m, Cuvier, Oss. Foss. pi. 102.
fig. ii. A 5) form the radial high and perpendicular border of the pig's unciform; at
the upper part of this perpendicular border abuts the os magnum, at the lower the
beak-like projection of the third metacarpal; this last projection, therefore, is more
horizontal in Suidee, and not so oblique as in Anoplotherium and ffyopotarnus. The
posterior projection of the unciforme is tolerably broad and well developed, as is seen

in Plate XXXVIII. fig. 7.

Dimensions.

Transverse breadth

Unciform Unciform

from Puy. from Hempstead (fig. 9).

. . 22 17

Depth, antero-posterior

. . *^^2T -*- •

A JL t/ lei il b . . . . e * i

. . 17 12

I had already completed this description when I received unexpectedly, through the
kindness of Mr. Davies, sen., of the British Museum, two carpal bones from the Museum
in Cambridge. On examination both proved to be unciforms. One of them (Plate
XXXVIII. fig. 9 & 9') corresponded very closely to our unciform from Puy, being,
however, much smaller; it may have belonged to some of the smaller Hyopotami,
whose metacarpals, corresponding entirely in shape with those from Puy, are often
found at Hempstead. This small unciform is figured in Plate XXXVIII. fig. 9, front
view, 9' the distal surface ; it is from the left side, like the corresponding one (figs. 5 & 7)
from Puy. Its distal articular surface (Plate XXXVIII. fig. 9 ; ) has three facets — one for
the beak-like projection of the third metacarpal, a central one for the fourth meta-
carpal, and an outer (v.) for the metacarpal of the small or fifth digit. The posterior
part of this small unciform is drawn out into a tolerably broad backward projection.

The second of the two unciforms received from Cambridge was found at Hordwell";
it proved a most valuable addition to my materials, as, after a careful study, I arrived at
the conclusion that it undoubtedly belonged to the large didactyle form referred by me
to the genus Diplopus with the specific name Aymardi. Figs. 8 & 8', Plate XXXVIII.
give a front and a distal view of this interesting bone. Comparing it with the same bone
from Puy, or the small one from Hempstead, we are at once struck by their difference.
The front view, in consequence of the large development of the inferior radial angle,
presents a much squarer outline ; the inner vertical wall of the bone has become much

* This makes the Hyopotamus unciform look exceedingly like that of a Palceotherium crassum ; but this
similarity is only superficial, as the distal surface and the posterior beak are very different from what we see
in Palaeotheridae.

MDCCCLXXIII. I

52 DK. W. KOWALEVSKY ON THE

higher, reminding one precisely of what is seen in Sus, and even more so in Hyomoschus.
The distal surface (Plate XXXVIII. fig. 8') is also very different. The facet for the
third metacarpal, instead of being oblique as in the bone figs. 5 & 7, % Plate XXXVIII.
(where it forms an oblique truncature of the inferior radial angle on the unciform), has
gone over to the inner or radial vertical wall of the bone, forming a facet homologous
to that of the Hog, composed of the facets k+moi the Anojplotherium (Cuviee, pi. 102.
fig. ii. A 5). In consequence of this the beak-like projection of the third metacarpal in
the two-toed form is more horizontal (Plate XXXVIII. fig. 6, in.) than in the four-toed
form (fig. 5, in.), approaching what is presented by the Suidce. The distal surface (fig. 8')
shows no trace of the outer facet for the fifth metacarpal, and is entirely taken up by
the greatly enlarged facet of the fourth ; and the margin of this facet is so sharp that it is
evident that the rudiment of the fifth metacarpal (whose existence is proved by a facet
on the outer side of the fourth metacarpal) did not touch the unciform, confining itself
to the lateral upper facet of the fourth metacarpal.

The posterior projection of the unciform (fig. 8') is very much broadened, so that the
whole distal surface acquires a somewhat square outline ; but what is more interesting,
the inner or radial inferior margin of this projection has a large round facet* by which
it articulated posteriorly with the os magnum. We never meet with such an enlargement
of the posterior par tof the unciform in the true Suidce , as in these the posterior and
inferior parts of the unciform and magnum do not articulate together. Such an enlarge-
ment, however, is seen in Dicotyles, where the unciform sometimes touches the magnum,
while in Hyomoschus we see, on the inner side of the same posterior enlargement of the
unciform, a round facet for the magnum ; this facet is also very characteristic of the
unciform of all true ruminants. The reason of this closer articulation between the
magnum and unciform seems to lie in the reduction of the lateral metacarpals, which
caused the carpal bones to concentrate more towards the central part of the carpus.

The proximal, or upper, surface of this unciform is also much broader and flatter than
in the four-toed form, both facets for the lunar and pyramidal being more on one plane.
These two facets are very nearly equal in size, not as in the Pig, where the lunar facet is
much longer and broader than the facet for the pyramidal.

Dimensions.

Unciform of
Dipl&pus, from Hordwell.

Transverse breadth ....... 21^

Antero-posterior depth ...... 21

Height 19

Tarsus of Diplopus and Hyopotamus.

The general structure of the tarsus in the Hyopotamidee is quite conformable to the
typical structure shown by all Paridigitata ; and, in considering the shape of each of its
constituent bones, I shall point out the features in which it resembles, or differs from,

* This facet is seen in %. 8 f from below, as a projection of the posterior and inner margin of the bone.

OSTEOLOGY OF THE HYOPOTAMDm 53

the other representatives of the same order. Fortunately the bones of the tarsus are
much more numerous than the carpal bones both at Puy and in England ; and I have
been enabled to restore it completely, with the exception of the first cuneiform, which
is wanting.

Calcaneum (Plate XXXV. fig. 4, Plate XXXVII. fig. 21, c). — This bone presents a
remarkably uniform shape through the whole range of Paridigitata ; in Hyopotamus
(Plate XXXVII. fig. 21, c) and Dvplqpus (Plate XXXV. fig. 4) it is very similar to the
calcaneum of a pig. Owing to this uniformity of the calcaneum in the Paridigitata, we
must extend our comparison to the other series, or Imparidigitata, if we intend to find
out what are really the characteristic and essential features of a Paridigitate calcaneum.

The general shape of the calcaneum in the entire order of Ungulata is, on a cursory
glance, very similar ; indeed, to find the typical differences we must enter into a deeper
analysis of the bone in question. If we put before us some calcanea belonging to
animals of the Imparidigitate series, for instance that of a Palwotherium, Tapir, Khino-
ceros, or Horse, and try to compare them with those figured in our plates, or, still
better, with the calcaneum of a Pig in naturd, we shall immediately see that in the
first * the upper and outer edge of the calcaneum, where it bends down to enlarge into
the sustentaculum, forms a large articular projection (fig. 4', e), which enters below
and behind the external pulley of the astralagus ; to the inner side of this articular
facet we have the sulcus sustentaculi (fig. 4', s), and, on the other side of this sulcus,
another large articular oblong internal facet for the astragalus (fig. 4', i). In corre-
spondence with this, the posterior part of the astragalus of all Imparidigitata presents
two large principal facets for the calcaneum — an outer facet, under the outer pulley,
and an inner oblong facet.

Looking now, with a view to a strict comparison, at the same upper edge of the cal-
caneum of a Paridigitate (Plate XXXV. fig. 4 ; Plate XXXVII. fig. 21), we see that,
at the point where the anterior edge bends down into the sustentaculum, it is divided
into two parts by the sulcus sustentaculi ; the external part, being the direct prolonga-
tion of the anterior edge, forms a prominent articular facet for the fibula (Plate XXXV.
fig. 4, ff) ; the inner part, being situated on the sustentaculum proper, gives a large
single facet for the posterior part of the astragalus (as). Considering these two facets
in reference to the sulcus sustentaculi, I thought that the outer astragalean facet of the
calcaneum of the Imparidigitata (e, fig. 4') ought to be homologized with the fibular
facet of the calcaneum of the Paridigitata (fig. 4, ff\ and the inner oblong astragalean
facet of the calcaneum of Imparidigitata (i, fig. 4') with the single astragalean facet on
the sustentaculum of Paridigitata. Comparing in naturd the calcaneum of a pig and
a horse, this is the view which seems most natural : it looks quite as if the outer astra-
galean facet of the Horse's calcaneum, instead of being buried under the outer pulley

* To make my comparisons better understood I figure the front view of an Imparidigitate calcaneum (Plate
XXXV. fig. 4') ; it belongs to Anchitherium. The calcaneum (Plate XXXV. fig. 4') is a left one, while the two
calcanea (Plate XXXV. Hg. 4 and Plate XXXVII. fig. 21) are right.

54 DE. W. KOWALEYSKY ON THE

of the astragalus, had got to the outside of it, and were articulated with the fibula. Such
a view of this homology is still more strengthened if we consider the relation between
the two bones in Anoplotherium (Plate XXXVII. fig. 11, and De Blaikyille, Anopl. iv.) ;
in this animal the fibular facet of the calcaneum is still covered by a process of the astra-
galus; it seems as if it were just one of the stages of the progressive march of the
fibular process from under the outer pulley to the outer side of the astragalus.

As all Imparidigitata have two chief astragalean facets on their calcaneum (Plate
XXXV. fig. 4', e and i), and no facet for the fibula, and as all Paridigitata having a
fibular facet on their calcaneum (fig. 4, ff) have only one chief astragalean facet, it
seems natural to say that the second (external) astragalean facet of the Imparidigitate
calcaneum is not lost in the Paridigitata, but has been made use of for the formation of
the fibular facet. But there seems to exist an objection to this view ; and it is furnished
by the Macrauchenia, the strange South- American form, which seems really to be a sort
of Imparidigitate Camel. The calcaneum of Macrauchenia, however, is not known ; but
the astragalus figured by Professor Owen in the ' Zoology of the Voyage of the Beagle '
(plate xiv. fig. 4) clearly shows that there were 'two principal facets for the calcaneum,
as in all typical Imparidigitata. Now the fibula has also on its distal extremity an
articular facet for the calcaneum* ; and this facet presupposes unerringly a similar fibular
process on the calcaneum, as it exists in the Paridigitata. If this should be really the
case, then the existence of two astragalean facets and a third fibular one will perhaps
invalidate the view as to the homology explained above, and it would in this case stand
thus: — that the single large astragalean facet in the calcaneum of the Paridigitata is
homologous to both facets of the Imparidigitata (tf+i, fig. 4'), the fibular facet of the
former being a new superadded character, not found in any of our living or fossil Impa-
ridigitata, but exhibited by the South-American Macrauchenia. So that, even with the
large material in our hands we cannot quite certainly determine the homology of different
parts of such an important bone as the calcaneum in both the chief series of Ungulata ;
and as we have no forms linking these two typically different calcanea together, this is
one proof more of the very ancient separation of this order into its two principal groups.

The calcaneum of Ilyopotamus (figured in \ nat. size, Plate XXXVIII. fig. 21) and of
Diplopus Aymardi (Plate XXXV. fig. 4, nat. size) resembles very much that of a pig,
with some slight differences. The sulcus which divides the fibular facet [ff) from the
astragalean is very deep ; this last facet has a slight angular rising or ridge nearer to
its inner border, which fits into a corresponding slight depression in the posterior
(calcaneal) surface of the astragalus ; besides, as in all Ungulata, there is, at the inferior
and inner border of the processus anterior, a smooth and large surface (Plate XXXV.
fig. 4, a s) by means of which the inner wall of the calcaneum fits to the outer border of
the inferior pulley of the astragalus. The fibular process (ff) is not so high, but much
deeper antero-posteriorly than in the Suina : and we have seen that the distal extremity
of the fibula is also extended in the same direction. The inner surface of the fibular

* Yoyage of the Beagle, p, 51.

OSTEOLOGY OF THE HYOPOTAMIDJE.

process is very smooth, without the large prominence we see in pigs, where it enters
deeply into the external wall of the astragalus. The inferior surface of the processus
anterior is occupied by one tolerably broad facet for the cuboid (Plate XXXV. fig. 4, cb).
On the posterior surface of the thickened back end is seen a groove for the tendon of
the musculus plantaris, which, as in all Ungulata, was certainly developed as a flexor
digitorum ; this groove is shallower than in the Suina.

Dimensions of Calcaneum.

Length

Height at the fibular facet . . .
Length of the processus anterior
Greatest breadth „

Hyopotccmus,

Diplopus,
from Hordwell

from Puy.

(Plate XXXV.
fig. 4).

23*

The astragalus (Plate XXXVII. fig. 21). — This is an extremely characteristic bone
for determining the natural affinitiesof Ungulata ; in Dvplopm it is very nearly like as
in Hyopotamus, and conforms to the general shape of this bone through the whole
range of Paridigitata. It has a double pulley — a proximal one, fitting the distal ends of
the tibia and fibula, and a distal pulley fitting the navicular and cuboid. In all Impari-
digitata the astragalus has only one pulley, on its upper or proximal end, while the lower
or distal is flattened, though articulating with the same two bones. (In Macrauchenia
and in the Horse, the distal surface of the astragalus, articulates only with the navicular
bone.) This character of the astragalus is one of the best by which to distinguish at a
glance the two series of Ungulata; and we know of no living or fossil animal which presents
an astragalus linking these two divisions together. True, in the astragalus of Hippopo-
tamus (Plate XXXVII. fig. 9, a), owing to the broadness and thickness of all the tarsal
bones, the inferior pulley is so exceedingly low that it reminds one of the same distal
surface of the astragalus in Bhinoceros (see De Blainville's plates) ; but this likeness
seems to be only an analogical one, both animals having heavy tarsal bones adapted
very nearly to the same function.

Although extremely like in general shape to the astragalus of a pig, that of the
Hyopotamus is proportionately much higher, the distal and proximal pulleys being divided
by a larger interval than in the Suidse. That part of the distal pulley which fits the
cuboid is perhaps produced a little more downwards than in the Suina, though the
difference is very slight. In comparing the large astragali of the didactyle Diplopus
with the smaller ones of Eyojpotamus, we find very few differences indeed; but it may
be mentioned that the proximal pulley of the two-toed form is relatively a little higher
and enters more deeply into the distal end of the tibia (Plate XXXVI. fig. '7), which may
be due to the greater narrowness of the reduced foot, requiring a more close articulation
with the tibia.

DB. W* KOWALEVSKY ON THE

Dimensions.

, ■■;..- ■.— ■-■ — .-— . — , ~— — , — — ?-— . — — — —

Hyopotamus.

Diplopus Aymardi.

Puy.

Hempstead.

Hordwell.

Length, tibial side

40 ,47
44|, 52
Ax. , £\j

24|, 30

44, 43, 40
50, 50, 45

25, 24, 23

26, 26, 23|

28,28

OOp

28i
26%
29

Length, fibular side

Anterior breadth in the middle

Transverse breadth, proximal pulley ....

The cuboid (Plate XXXVIII. figs. 1, 3, 10, 11). — I have examined numerous and well-
preserved specimens of this bone from Puy, Hempstead, and Hordwell ; the cuboids
from Puy and Hempstead (figs, 1,10) clearly indicate the four-toed form called Hyopo-
tamus, while the bones from Hordwell (figs. 3, 11) belonging to the genus called by me
Diplopus, indicate, in the most unmistakable manner, that in this last animal the foot
was didactyle, as the inferior surface of the cuboid shows no trace of a facet for a fifth
digit as in Hyopotamus.

As the cuboid is one of the most important bones of the tarsus, I will give a detailed
description of it, with as many figures as I can afford ; but should the reader take a
cuboid of a hog in naturd, then many points that are difficult to explain by words will
at once be plain to him.

I will begin, as I have nearly always done, with the description of the cuboid of
Hyopotamus (Plate XXXVIII. figs. 1, 10), and afterwards compare it with that of the
Diplopus (figs. 3, 11). As in all Paridigitata, which have the navicular and cuboid
distinct, the front face of the cuboid of Hyopotamus is cut out en equerre (Cuv.) ; this is
brought about by the circumstance that the cuboid articulates with two bones of the
first row of the tarsus, the calcaneum and astragalus ; and the facets for these two
bone® lie in different levels, the rounded crescentic facet for the outer half of the distal
pulley of the astragalus rising in front, while the facet for the processus anterior calcanei
presents a flat surface sloping forwards, as is plainly to be seen in the figures (fig. 1 cl
and as). The upper or proximal surface of the cuboid is occupied entirely by these two
facets, one internal, shaped like a concave crescent for the astragalus, the other external,
like a convex sloping surface for the calcaneum. The proximal surface of a hog's cuboid
shows the same facets ; only the external one is often cut transversely in two by a rough
deep notch for the passage of vessels and ligaments : this division of the astragalean
facet of the cuboid is nearly constant in the Hog and Dicotyles, though I did not find
it in Phacochoerus and JBabirussa. The relations between the transverse breadth of the
two proximal facets vary in different genera. In Anoploiherium, the calcaneal facet is
much broader than the astragalean (Plate XXXVII. fig. 11) ; but in Hippopotamus and
Hyopotamus the relation is reversed, while in most Suidee both facets are very nearly of
equal breadth.

The internal wall of the cuboid shows on its upper margin a small facet for the
neighbouring navicular, and a little lower a very small rising which corresponds to the

OSTEOLOGY OF THE HYOPOTAMIDiE. 57

nterval between the distal surface of the navicular and proximal of the great or third
cuneiform ; the rising was very slight on our specimen, and is only imperfectly seen in
the drawing. On the posterior part of the inner wall we see two facets divided by a groove
for the external and posterior part of the navicular (in the Hog these two facets have
coalesced into an elongated single facet on the posterior border of the inner wall of the
cuboid). This posterior margin of the internal wall shows in Hyopotamus a considerable
bulging inwards, into the middle of the tarsus (seen in Plate XXXVIII. fig. 1 0), which
makes this cuboid much broader than the corresponding bone of the Diplopus (fig. 11),
where this bulging is much less, nearly identical with that of a hog's cuboid.

The outer surface of the cuboid is rounded. If we look at the external surface of the
cuboid of a hog we shall see a deep sulcus for the peronaeus tendon situated between the
posterior prolongation of the cuboid and the distal articular facet ; this sulcus is absent in
Hyopotamus, owing to the fact that the posterior part of the cuboid is not produced
downwards, as we shall immediately see. The posterior surface of the cuboid is con-
siderably broader than the anterior, as is to be seen in fig. 10, while in Diplopus (fig. 11),
and especially in the Hog, the converse is the case. Besides, looking at the cuboid from
the posterior aspect, we perceive a very broad and rough transverse ridge for muscular
and ligamentous attachment, running through the whole breadth of the bone as it is seen
in fig. 10 from below. The Anoplotherium and the Hippopotamus have nearly the same
transverse ridge on the back part of the cuboid ; only it descends lower down, while
in Hyopotamus this ridge does not reach the level of the distal articular surface of the
cuboid, which is the lowest point of the bone; we shall by-and-by indicate the difference
presented in this respect by the cuboid of Diplopus.

The distal surface (Plate XXXVIII. fig. 10) is the most important, as it presents the
articular facets for the two outer metatarsals. Not only in Ungulates but in all recent
and fossil Mammalia the cuboid of the tarsus, as well as its homologue (the unciform) in
the carpus, gives an attachment to the two outer metatarsals and metacarpals of the
.foot (the fourth and fifth) if they are not entirely reduced. In looking at the distal
surface of the cuboid in Hippopotamus ', we shall perceive a large central facet for the
fourth metatarsal, and outwards from this another good-sized facet for the well-developed
fifth metatarsal. In the common Hog, nearly the whole distal surface of the cuboid is
taken up by the much developed fourth metatarsal; but still there exists at the
outer part of this large facet a smaller and oblique facet for the still functionally developed
fifth digit. In Dicotyles, where the fifth digit is lost, or only a short rudimental
metatarsal of it remains, the whole distal surface of the cuboid is taken up by the
fourth metatarsal, the rudiment of the fifth being only attached to the outer? side of the
fourth metatarsal, without touching the cuboid (De Blainville, 8us 9 plate vii. Dicot
laMatus, lowest left figure). In the Anoplotherium commune, as well as in A. tridactylum,
there is not the slightest trace of a facet for a rudimentary fifth digit ; but in Xiphodon,
which also has only two developed metatarsals, we have on the distal surface of the
cuboid a large facet for the fourth metatarsal, and an outer small facet for the rudi-
ment of the fifth, which certainly existed, as the outer border of the fourth metatarsal

58 DE. W. KOWALEVSKY ON THE

has also a small facet, to which this rudiment was attached (Cuviek does not mention
this, as he treats both Dichobune and Xiphodon very briefly). This additional facet
for the rudiment of the fifth metatarsal is to be seen in De Blainville {AnopL
plate v.) ; it certainly existed in Dichobune and Cainotherium, as both have a completely
developed fifth digit in the pes and nianus.

Now, looking at the distal surface of the cuboid of Hyopotamus (Plate XXXVIII.
fig. 10), we clearly perceive two large articular facets (iv. & v.) nearly in the same mutual
relation as in Hippopotamus ; the two facets are numbered iv. and v. in correspondence
with the metatarsals they support ; and their outline corresponds very nearly with the
proximal surfaces of these two metatarsals (Plate XXXVIII. fig. 1, iv., v.). The inner
border of the facet iv. is convex, the outer (near the ridge that separates the two facets)
concave ; and corresponding to this the inner border of the fourth metatarsal is con-
cave, and the outer convex, to fit the cuboid. The facet v. is a little convex and tri-
angular ; and the metatarsal v. is shaped in a corresponding manner.

Besides the cuboid figured from Puy, I had a smaller cuboid from Hempstead, with
two facets on its distal surface for the fourth and fifth metatarsals, and another from
the same locality much larger than the Puy specimen, but corresponding closely with it ;
unhappily this last large specimen is much rolled, and the articular facets are not quite
distinct.

If we compare with the cuboid just described the same bone. from a nearly complete
foot found at Hordwell (Plate XXXVIII. figs. 8, 4, and, distal view, 11) belonging to the
Diplopus, or the two-toed form of this family, we shall see important differences which
must clear away any possible doubt as to the number of digits in the genus Diplopus.
The front view of the cuboid, fig. 3, shows it to be a little narrower than the same bone
in its smaller but four-toed congener ; and this is especially the case if we compare the
posterior parts of both bones, as seen in figs. 10 and 11. On the inner wall, the middle
anterior rising which enters into the interval between the navicular and third cuneiform
is more distinctly developed, as seen in fig. 3 ; the rising, with a facet for the distal surface
of the navicular, on the posterior border of the cuboid, is not so thick and not produced
so much inwards as in Hyopotamus, but repeats entirely the arrangement seen in the Hog.

The external surface is more hog-like in appearance, and differs from the same
surface of Hyopotamus by the presence of a deep sulcus (fig. 11) between the distal
articular facet and the posterior prolongation (fig. \\,fmy\ which, in this genus, is
produced much lower down than in Hyopotamus. Instead of the broad transverse ridge
seen on the posterior surface of the cuboid in the Hyopotamus (fig. 10, tr), the cuboid of
the two-toed Diplopus has this ridge prolonged downwards in a beak-like process quite
of the same shape as in the common Hog (fig. 4, h.cb). This posterior beak descends
lower down than the distal articular surface of the cuboid, and exhibits on its inner side an
elongated facet (fig. 11, f.my) by which this beak articulates with a corresponding cuboid
facet on the outer side of the posterior prolongation of the fourth metatarsal (fig. 4", iv.
cb.f). This articulation is observable, though imperfectly, in fig. 4, where the beak of
the cuboid (bxb) is seen descending on the other side of the fourth metatarsal. In my

OSTEOLOGY OF THE HYOPOTAMIMl.

specimens of Hyopotamus from Puy the posterior prolongation of the fourth metatarsal
is not well preserved (Plate XXXVIII. fig. 2', iv.) ; but as there is no downward pro-
longation in the cuboid and no facet, the cuboid seems not to have articulated with this
posterior prolongation of the fourth metatarsal, and it does not so articulate in Anoplo-
therium and Hippopotamus. In the Hog the relations are just the same as in the
didactyle Diplopus ; only this beak has on its outer side another small facet, for the
posterior prolongation of the outer ox fifth metatarsal — which is wanting in Diplopus^
and, though present in Hyopotamus^ has no such posterior prolongation as in the Hog.
A glance at a hog's tarsus will make all this much clearer.

The distal surface of the cuboid in Diplopus (Plate XXXVIII. fig. 11) is entirely taken
up by a single large facet for the fourth metatarsal, leaving no place at all for an additional
digit as in fig. 10, though a rudiment of such fifth digit certainly existed, as is proved by
a small facet on the outer border of the fourth metatarsal, to which this rudiment was
undoubtedly attached, without ascending, however, so far as to touch the cuboid.

All these differences clearly indicate that the Hyopotamoid form to which this cuboid
(fig. 11) belonged had no fifth metatarsal, but only a rudiment of it ; we have seen the
same thing in the manus, where the unciform, instead of supporting two toes, is shown
by the distal articular facets to have supported only one (the fourth) while a rudiment
of the fifth, indicated by an external facet of the upper border of the fourth metacarpal,
did not reach the unciform and left no trace upon it.

Dimensions of the Cuboid.

Height, anterior

Breadth, upper part

Breadth, inferior, on the distal surface
Antero-posterior depth

Hyopotamus,

from Puy

(Plate XXXVIII.

figs. 1, 10).

Hyopotamus,
from Hempstead.

Diplopus,

from Hordwell

(Plate XXXVIII,

figs. 3, 11).

26
21
18
25

19, 31
17, 24
15,21
25, 28*

23
19

16i
25

The navicular (Plate XXXVIII. figs. 1-4, n). — This bone is nearly identical in both
genera, and has the same general shape as in all Paridigitata, in which the navicular is
separate from the cuboid. The upper or proximal surface is entirely the same as in the
Hog, except that the middle rising which fits into the groove of the distal pulley of the
astragalus is more rounded. The inner (fibular) surface has, on the anterior and
posterior part, facets corresponding with those already described on the tibial side of
the cuboid. The posterior termination is rounded and only moderately produced
downwards (as seen in figs. 2 and 4), without such large beak-like prolongation as is
seen on the posterior side of the navicular of a hog (Plate XXXVII. fig. 12, n) or
JDicotyles (fig. 13. n).

The distal surface is the most important one, as it supports the three cuneiform bones,
which carry the digits. All the anterior part of the distal surface of the navicular is

MDCCCLXXIII. K

DK. W. KOWALEVSKY ON THE

occupied by a broad facet for the great or third cuneiform, which supports the third
metatarsal. I have not figured the distal extemity of the navicular ; but, by looking
at the distal surface of the Anoplotherium navicular (Plate XXXVII. fig. 17) or that of
a hog, the reader may form a correct idea of it. This facet is much narrowed in its
posterior part, and then expands slightly again, forming a small facet for the second
cuneiform ; behind this there is a slight sharp ridge, which is seen tolerably well in
the side view, fig. 4, Plate XXXVIII., as well as on a navicular of the Hog (Plate
XXXVII. figs. 12, 13, n); behind this ridge is the facet for the first cuneiform. This
last facet is concave on some of the naviculars of Hyopotamus, and convex on the two
others belonging undoubtedly to the two-toed Dijplojpus ; whether the difference was
constant I cannot tell without ampler materials. I find no essential differences in the
shape of the navicular bone in the two genera ; and my description, made from good
specimens of the Diplopus, may apply to the Hyopotamus also ; of this last I had only
one, not quite perfect, specimen of navicular from Puy.

Navicular of the Tarsus.

• • * • *

•••• *••»

Anterior height . . ......

Breadth, proximal face

Depth, antero-posterior »

Breadth, distal surface

Hyopotamus.
from Puy.

24
14

Diplopus,
from Hordwell.

12|, 14
15, 17
26, 30
14!, 20

The cuneiforms (Plate XXXVIII. figs. 1-4, c 3 ). — These are small but very important
bones, on account of their relation to the digits which they have to support. The second
row of the tarsus, containing the three cuneiforms and the cuboid, exhibits very clearly
the complete homology of the fore and hind limbs. As in the complete manus we have
in the second row of the carpus three distinct bones (the trapezium, trapezoideum, and
magnum), each supporting a separate toe, and one external bone, the unciforme,
giving support to the two outer digits (the fourth and fifth), so in the pes we perceive
also an entirely homologous set of three cuneiforms (first, second, and third), each
supporting a separate toe, and one external bone, the cuboid, giving support to the two
outer digits of the pes (the fourth and fifth). Where all the five digits are completely
developed this rule admits of no exception whatever in the whole range of Mammalia ; but
where, as in Ungulata, the number of the digits becomes sometimes greatly reduced, the
relation of the carpal and tarsal bones to the remaining digits is slightly changed, though
all the three cuneiforms are still retained, and I know of no instance where one of
them is wanting. This persistence of cuneiforms is a strange fact ; for in the carpus,
which is generally less reduced than the tarsus, the trapezium (homologue of the
first cuneiform) is certainly lost in Horses and Ruminants, while its homologue in the
tarsus is still retained in a recognizable state. These relations, slight as they may
seem, are important ; we must know all the details and the true homology of the

OSTEOLOGY OE THE HYOPOTAMIDJ2. 61

smallest bones to establish clear lines of descent. I shall be obliged to discuss the
question of the cuneiforms again when I come to describe the metatarsals, and
will confine myself in this place to a general review of them in the Paridigitata, and
a special description of these bones in Hyopotamus and Diplopus. As the Hippo-
potamus is the most complete of the living Paridigitata, we see in the side view
of its tarsus (Plate XXXVII. fig. 10) all the three bones (c\ c 2 , c d ) fully developed;
but as the first digit is aborted, its typical tarsal bone, the first cuneiform (0J, gives
additional support to the second digit. Each of the two inner digits (the third
and second) are supported by a separate cuneiform. We see besides, in this side view
of the tarsus, that the second digit (fig. 10, il), besides its own tarsal bone (the second
cuneiform), seeks additional support from the large third cuneiform : we shall discuss this
relation when we come to the metatarsals ; and I will merely observe here that it seems
to be a very ancient one, which existed perhaps in the progenitor of the whole class of
Mammalia. In the Hog (Plate XXXVII. fig. 12) we find the three cuneiforms
quite distinct ; but there is a great change in their relative position : instead of the
third cuneiform giving additional support to the second digit, it is now not only
entirely occupied by the enlarged third digit, but even the greater part of the second
cuneiform, leaves the reduced second digit to give additional support to the enlarged third.

The fir st cuneiform is also peculiar in the Hog ; it is thrust like a wedge between
the posterior beak-like prolongation of the navicular, the head of the second metatar-
sal, and a posterior process of the third metatarsal, to which it gives an articular facet*
Eemembering what we have said in reference to an analogous beak-like prolongation
of the cuboid in hogs, and looking at a hog s tarsus in naturd, we shall see that the
wedge-shaped first cuneiform, together with the cuboid, press from within and
without on the posterior processes of the third and fourth metatarsals, and make the
two middle metatarsals act like a single undivided bone : this is moreover aided by
a special provision in the articulation of the distal extremities of these metatarsals
with the first phalanges ; namely, the outer borders of the combined distal extremities of
the two metatarsals are produced down less than the inner, so that the digits, being
expanded in treading on the ground, tend to meet by their proximal ends and compress
both metatarsals firmly together, thus materially aiding the compression of their upper
extremities by the beak-like prolongation of the cuboid and the wedge-like first
cuneiform. There is nothing of the sort in Hippopotamus ; but the same provision
existed in the two-toed Diplopus.

JDicotyles (Plate XXXVII. fig. 13) presents nearly the same relations ; only, the second
cuneiform (c 2 ) has gone entirely over to the third digit; and after this there was no
assignable reason why these i^joo bones, giving support to a single metatarsal, should
not coalesce : and this they have done in Hyomoschus and Tragulus (figs. 14 and 15,
c 3 , c 2 ), where the navicular has also coalesced with the two cuneiforms, the first cuneiform
(c l ) remaining entirely distinct in both. In a Miocene ruminant from Auvergne,
fig. 16 (Amphitragulus 1), with a rudimentary second metatarsal, we may also clearly

K 2

62 DE. W. KOWALEVSKY ON THE

perceive that the large cuneiform is formed by the coalescence of the second and third,
while the first cuneiform remains distinct, giving support to the rudimentary metatarsal n.
The same relation may be seen in our recent Ruminantia, whose large cuneiform bone is
therefore to be taken as the homologue of the third and second, the first remaining distinct.

Having traced our three cuneiforms, even in the most reduced Paridigitata, it is
natural to ask how the matter stands in Anoplotherium, the reputed prototype of every
Paridigitate. In this respect the confusion is very great ; and it originated with Cuviee.
In his description of the Anoplotherian pes he says (' Oss. Foss.' 4th ed. p. 147,
plate 128), that he found " le grand cuneiforme C " (which is our third cuneiform)
and, besides, another bone (plate cxxviii. fig. 2, h), which he calls " osselet surnu-
meraire," and assumes to. be a rudiment of 'the second digit, saying (p. 147) that it
articulates to the facet i of the scaphoid, and the facet Jc of the third metatarsal.
(This bone is, no doubt, our first cuneiform, Plate XXXVII. fig. 11, c x .) If we
look at the restoration of the pes of Anqplotherium in De Blainville, and read his text,
we shall see that he speaks of three cuneiforms, and even figures them (De Blainville,
'Osteogr.' Anopl. plate iv., side view of pes); B is the third cuneiform, C the second;
and the bone lying backwards from it is certainly the first cuneiform. I take this
restoration to be in the main right ; only the second cuneiform, which is represented in,
the side view as articulated to the scaphoid, has, in his figure of the distal extremity
of the scaphoid (a little below), no such facet on this bone, the facets figured being for
the first and third cuneiform only. But if we read his text we shall find the con-
fusion which is so characteristic of his palseontological writings. He says : — " Les
os cuneiformes sont au nombre de trois: le premier, assez allonge .... colle contre
le metatarse ; le second presque de meme forme, un peu moins recule, articule avec
le scaphoide d'une part et le metatarsien du medius de 1' autre, et enfin le troisieme
.... articule carrement avec le scaphoide en haut et la metatarsien du medius en
bas."— AnopL p. 35.

So, in the interpretation of De Blaustville, the first cuneiform does not touch the
scaphoid at all (in the description of the scaphoid, p. 35, he says : — " le scaphoide a
a son bord interne un tubercule ovale un peu saillant pour l'articulation du second
cuneiforme, outre un beaucoup plus grand pour le troisieme "), but is applied to the
third metatarsal, the second cuneiform being, in his description, also articulated with
the same metatarsal ; and the third cuneiform is the chief support of the third digit.
All the three cuneiforms, then, are said by De Blainville to articulate with the third
metatarsal ; but I regret to say that this is entirely incorrect. I enter into such detailed
descriptions only because such matters should be at once set right ; the real state of
things is this : — The Anoplotherium possessed all the three typical cuneiforms * in their

* Their relations to the metatarsals were entirely identical with what we see in Hippopotamus : the third
cuneiform supported the third digit ; the second cuneiform supported (the rudiment of) the second digit ; and
the first cuneiform was articulated to the navicular above the second cuneiform in front, and the rudiment
of the second digit below.

OSTEOLOGY OF THE HYOPOTAMID^. 63

typical relation ; and, besides, a rudiment of the second digit pressed against the upper
margin of the third digit. To prove this, I have figured the pes of the Anoplotherium
from Vaucluse, where this rudimentary digit is completely developed (Plate XXXVII.
fig. 11), we see in this foot the three typical cuneiforms nearly in the same relation as
in Hippopotamus ; the third cuneiform articulated to the third metatarsal, the second
cuneiform to the metatarsal of the short second digit ; the first cuneiform, seen behind,
aids to support this short second digit. I have represented in Plate XXXVII. fig. 19
the distal surface of the navicular of this Anoplotherium tridactylum ; and we distinctly
see the three facets for the three cuneiforms, which all articulated with the navicular
bone. Together with these navicular bones from the tridactyle form, we have from
Vaucluse some naviculars of the didactyle species (fig. 18) : the complete second digit is
reduced in this species to a mere rudiment; and we see a corresponding diminution of the
facet for the second cuneiform which supported this rudiment, the facet for the third
remaining nearly of the same size. Now in the Anoplotherium from the Paris gypsum,
as I see by a navicular in the British Museum (fig. 17), this facet for the second cunei-
form is even smaller, but it exists still. As far as I can judge by a cast of another
Anoplotherian pes in the Museum, the facet for the second cuneiform is absent from
it ; but still we see a trace where this very reduced second cuneiform leaned against the
third, perhaps not touching the navicular; the posterior navicular facet for the first
cuneiform remains very large. The reduction of metacarpal and metatarsal bones
always precedes that of the carpals and tarsals ; and when a rudiment of the former still
lingers, we may be sure to find the corresponding tarsal or carpal ; and so it is in
Anoplotherium, which possessed besides its three typical cuneiforms a rudiment of the
second metatarsal.

Proceeding now to the Hyopotamus, we find that it, as well as the Diplopus, had
three cuneiforms to its tarsus; but unfortunately the first is absent from all collec-
tions examined by me; the second is preserved, thanks to its tendency to coalesce
with the third or great cuneiform. The shape of the third cuneiform is very like
that of the same bone in the Hog or Anoplotherium ; it is articulated to the third
metatarsal below (Plate XXXVIII. figs. 1-4) ; and its tibial or inner border gives additional
support to the second metatarsal *, as its homologue, the magnum, does in the carpus.
The second cuneiform is a small nearly cubic bone. I have never seen it separately, as it
is always lost owing to its small size ; but it is represented in fig. 2, having coalesced with
the third cuneiform, and giving support to the second metatarsal. The third cuneiform
of Diplopus, figured in the nearly complete tarsus (Plate XXXVIII. figs. 3 and 4), is in
all respects similar to that of Hyopotamus ; only, as seen in fig. 4, it did not give any
attachment to the rudiment of the second metatarsal, as it does to the complete
second metatarsal in Hyopotamus; nay, more, the tibial upper margin of the third

* In fig. 2, Plate XXXYIII., the second metatarsal rises a little higher than the third, and touches the
internal side of the third cuneiform. The same is to be seen in Hippopotamus (Plate XXXYII. fig. 10) and
Anihracotherium. This relation is essential in an unreduced tarsus.

64 DE. W. KOWALEVSKY ON THE

metatarsal in this specimen has a very slight truncature or oblique facet, by means of
which this third metatarsal touched the second cuneiform, a deviation from typical
relations which is enormously developed in the Hog (Plate XXXVII. fig. 12, in. c 2 ) ;
the proximal surface of the rudimental second metatarsal (Plate XXXVIII. fig. 4, n. r)
is therefore a little lowered in this specimen, to allow the second cuneiform to touch the
third metatarsal. On the contrary, in two other specimens not figured, the relations
remain true to type, and it is the third cuneiform which gives a small facet to the
rudiment of the second metatarsal,

I have found, moreover, in the British Museum a specimen of the, Diplopus from
Hordwell in which the two cuneiforms (2 + 3) had coalesced in the same manner
as the two cuneiforms of the Hyopotarnus of fig. 2. The distal surface of these two
coalesced cuneiforms is given in outline below fig. 4, Plate XXXVIII., to show its cor-
respondence with the proximal surface of the third metatarsal ; and it answers so exactly
to the third metatarsal with the coalesced rudiment of the second, figured in fig. 4",
that one may think they belonged to the same individual [my coalesced cuneiforms,
however, are from the other (right) side] ; and the correspondence to the figured foot is
so exact that even the second cuneiform presents a small facet (the shaded bit of the
outline below fig. 4) to the third metatarsal. So that in this specimen the overdeveloped
third metatarsal seeks additional support from an adjoining tarsal bone. In the Suina
it accomplished the passage long before the loss of the second lateral digit, as seen in
figs. 12 and 13, Plate XXXVII.

A similar case is described in my Memoir on Anchitherium * : here also the third
cuneiform became, in the same way, confluent with the second, to pave the way fox
the overdeveloped third (and single) metatarsal to enlarge and usurp the second cunei-
form, which bone typically belongs to the second digit.

The first cuneiform is absent ; and, judging from the absence of a facet for this bone on
the posterior prolongation of the third metatarsal in Hyojpotamus (Plate XXXVIII.
fig. 2), we may infer that it did not articulate with it; but in Diplopus such facet
on the posterior prolongation of the third metatarsal is clearly seen (fig. 4", fcj,
and furnishes a conclusive proof that the first cuneiform was articulated to it in the
same manner as in the Hog. Above, it was articulated to the somewhat convex pos-
terior facet of the navicular (seen in figs. 4 and 2, Plate XXXVIII.), then touched the
back part of the small second cuneiform, and, being prolonged lower down, leaned
against the posterior part of the second metatarsal (as proved by a facet on this bone)
in Hyopotamus (fig. 2), or against the rudiment of this second metatarsal in JDiplopus, as
seen in fig. 4, Plate XXXVIII. In this latter the first cuneiform was probably thrust
like a wedge between this rudiment and the posterior prolongation of the third meta-
tarsal, in the same way as it is to be seen on a hog's tarsus (Plate XXXVII. fig. 12).

Metacarpus and metatarsus. — When I began to study the structure of the anterior
and posterior extremities of the Anthracotheridae and Hyopotamidse in the collections of

* Memoires de 1'Academie de St. Petersbourg, 1873.

OSTEOLOGY OF THE HTOPOTAMID^E. 65

the Museums of Lausanne and Puy I had no doubt whatever that the genus imperfectly
known under the names of Hyopotamm and Bothriodon had four completely developed
digits on both extremities ; and all I had seen in the collections of Puy confirmed this
inference; besides, as I could not find any difference in the dental or osteological
characters between these two supposed genera, I was obliged to unite them, giving
priority to Professor Owen's denomination. But, whilst studying the fossil remains
from Hordwell and Hempstead in the collection of the British Museum, I gradually
acquired the conviction that we have, in the English deposits, two very different forms of
the same natural family, which, though entirely similar in most of their osteological
characters presented a wide difference in the composition of their manus and pes,
a difference somewhat similar to the difference we see in our own times between
Hyomoschus and the true Euminants. One form had clearly four completely developed
digits, while the other had only two. At first I struggled against this conclusion, as the
fact seemed too singular, considering the similarity of all the other bones of the skeleton ;
but by-and-by, as I got more materials, all doubt cleared away ; and having once ascer-
tained this point, I felt obliged to subdivide the family of Hyopotamidae into two
groups or genera, Diplopus and Hyopotamus, whose chief distinction lay in the number
of digits. It is more than probable that each group or genus was represented by
several species, though this specific distinction is very difficult to establish. But I shall
return to this question at the end of my memoir ; at present I proceed with the descrip-
tion of the bones forming the metacarpus and metatarsus in both genera. As I have
generally done, I commence with the tetradactyle form, or the ffyojpotamus, and shall
proceed afterwards to the didactyle, or Diplopus.

Before, however, proceeding with the Hyopotamidae, it seems necessary to cast a
glance at the composition of the manus and pes in all Paridigitata, living and fossil,
as such a general review will throw more light on the importance and value of the
distinctions found in our fossil Hyopotamidee than a tedious and minute description
of their bones could do. Studying in detail the structure of the extremities of the
living and extinct Ungulata Paridigitata, and especially their metacarpals and meta-
tarsals, we meet everywhere, through all the great diversity of their forms, such striking
similarity in all the smallest details, such adaptation of one typical structure to widely
different conditions, that there seems to be no possibility of explaining it otherwise
than by a common descent (with modification and adaptation) from some single ancient
form which presented the arrangement we call typical for the whole of the Paridigitata.
Looking upon the whole community of Paridigitata as modified descendants of one
ancestral form, the theory of evolution says that each form must have inherited the
typical structure of the common ancestor, modifying it in each particular case to the
condition of its own life ; and though, after the lapse of a great period of time, the
diversity acquired by different forms would be prodigious, still every member of the cycle,
even at the furthest points of descending lines radiating from the common ancestor, would
by every single bone testify its relation to this common ancestor. And such is really the

66 DE. W. KOWALEVSKY ON THE

case with our recent Paridigitata. If the Hippopotamus and Phacochcerus, Tragulus and
Antilope, seem wonderfully different as being on the furthest points of the radiating
lines, still they have every one of them strikingly similar typical characters inherited
from the ancestral form which was the common progenitor of them all.

The chief and most obvious characters of a Paridigitate foot are to be seen in the
shape and connexions of its two middle fingers ; therefore I will commence with these,
and state some general features common to all Paridigitata, living or fossil, without
exception.

While, in the Imparidigitata, the axis in relation to which the whole foot is shaped is
given by the middle or third digit, the axial line passing through its centre, we have in
a Paridigitate foot always two middle digits (the third and fourth), which form the
centre in relation to which the foot is arranged, the axial line passing in the interval
between these two digits. These, then, may be taken as the principal digits of the
manus and pes ; they exhibit nearly always a certain mutual symmetry, and are inter-
locked at their upper extremities by a peculiar arrangement which is common to all
fossil and living Paridigitata, as far as we know them now.

The mutual interlocking of the middle metacarpals is effected in this way. The
fourth metacarpal has always, at its upper radial end, a process, or a smooth salient
margin, uninterruptedly connected with its proximal articular surface : this process enters
deeply into a corresponding excavation of the third metacarpal ; and this excavation
is overarched by a special ulnar process of the third metacarpal, which is inclined
outwards, and abuts against the unciform. This is to be seen on all the figures of
the carpus (Plates XXXVII. and XXXVIII.), and better still in De Blainville's Plates.
In consequence of this projection of the third metacarpal the fourth stands always a
little lower than its neighbour, and is supported entirely by the unciform, while the
third is supported by the os magnum and partly by the unciform (Plate XXXVII.
figs. 1, 2, 3, 20 ; Plate XXXVIII. figs. 5, 6), upon which it glides by means of its oblique
outer and upper process.

The two middle metatarsals (the third and fourth) (Plate XXXVIII. figs. 1 & 3)
show us something quite analogous ; the fourth gives off from its upper tibial side a pro-
minent process (not uninterruptedly connected with the proximal surface as on the fourth
metacarpal), which enters into a deep corresponding cavity in the outer side of the third
metatarsal. The fourth metatarsal is supported by the cuboid, the third by the third
cuneiform.

Such is the general rule of the interlocking of the two middle metacarpals and meta-
tarsals in all living and extinct Paridigitata : there is not a single exception to it ; and
the old Eocene Dichobune, as well as the recent Hippopotamus, present us with the same
relation. But it may be urged against me that such a relation is certainly not to be
seen in the recent Euminantia, whose middle digits have coalesced into one single cannon
bone. However, I think that the rule is even applicable to them. Although in the
recent geological period the Ungulata Paridigitata, generally, acquired an exceedingly

OSTEOLOGY OF THE HYOPOTAMIME. 67

reduced skeleton, we still possess nearly all the intermediate stages between four com-
pletely developed and distinct metacarpals and metatarsals, and their reduction to two
coalesced middle ones. Besides, the links which are wanting in the fossil state are often
furnished by the living forms in the course of their individual development before they
reach the adult age. Such links are presented by the Tragulidee and the posterior extre-
mities of Dicotyles and Hyomoschus. If we examine the bones of their feet at an early age,
when the fourth and third digits are yet distinct, we shall see exactly the same relation
as is laid down in our general rule. But even after the complete coalescence of these
two metacarpals and metatarsals, some traces of the original disposition are to be detected.
Examining the fore "cannon'' of the typical Euminantia (especially Deer) we shall see
that one half of the proximal surface, answering to the third metacarpal, is a little higher
than the other half, and always presents a produced ulnar margin, by means of which
it abuts against the unciform ; the same is to a certain extent seen in the back cannon,
where the proximal surface for the cuboid is often a little lower than the surface for
the third cuneiform.

These two middle digits (third and fourth) form the chief basis of the foot of a Pari-
digitate Ungulate ; to these are added, on the inner and outer sides, one digit more, the
second and the fifth. But so much is the skeleton reduced in the existing Paridigitata,
that there exists at present only a single form (the Hippopotamus), represented by a
single species (or two, as Chceropotamus liberiensis appears to be distinct), in which these
two lateral digits have a true functional importance ; and although the lateral digits still
exist in all the Suidae, Tragulidae, and Hyomoschus, they have no real functional import-
ance. Even in the Suidae the two middle digits are so largely developed in comparison
with the lateral ones, and all the bones of the carpus and tarsus have been so completely
taken for the use of these two middle digits, that we shall be guilty of no exaggeration
in stating that the Suidae might lose the lateral digits without their locomotion being at
all impaired. We even witness this process going on in Dicotyles — in which the lateral
digits are still more reduced than in the true Suina, and they begin to disappear altogether,
beginning at the external metatarsal If the lateral digits are still retained in the Suidae,
it is chiefly owing to the fact that the Hogs live generally in marshy places and on
muddy river-banks, where a broad foot is of great importance for not allowing them to
sink deeply into the mud. But if, by some geological change, their habitat should be
transformed into dry grassy plains, there can be no reasonable doubt that they would as
readily lose their lateral digits as the Palaeotheroids have lost theirs (perhaps by an
analogous change of habitat) in becoming transformed into the monodactyle Horse.
Should this occur only in a limited locality, and under circumstances admitting of no
migration (for instance, on a large island), it is very possible that the Suidse of this
particular island might lose their lateral toes, while others, which continued to live
under the old conditions, would retain theirs. If this should really occur, we should
have two groups of the same family, quite similar as to the large bones of the
skeleton, but dissimilar as to the structure of their feet, which is just what we witness

MDCCCLXXIII. L

68 BE. W. KOWALEVSKY ON THE

in the two groups of the didactyle and tetradactyle Hyopotamidse of the oldest Miocene
and Upper Eocene.

What, then, are the relations of these lateral digits to the whole manus and pes,
considered in their primitive unreduced condition. The inner, or the second lateral
digit of the manus is supported by the trapezoid ; it is not, however, limited only
to this bone, but leans a little over the inner edge of the third metacarpal, and goes to
touch the os magnum, which presents a special facet to this second metacarpal. We
may see this typical relation in the second digit of the Anoplotherium tridactylum
(Plate XXXVII. fig. 2), in the rudiment of the second metacarpal in the Paris Anoplo-
therium from the gypsum (see De Blainville, AnopL), and in the very complete fore
foot of Hippopotamus and Hyopotamus (Plate XXXVII. figs. 1 & 20, n. £, m). In the
posterior limb, the inner or second metatarsal is supported entirely by the second cunei-
form (the homologue of the trapezoid), and by the tibial edge of the great or third
cuneiform (the homologue of the os magnum), as maybe seen in Hippopotamus, figs. 9 & 10,
Anoplotherium, fig. 11, Hyopotamus (Plate XXXVIII. fig. 2), and Anthracotherium* '.

The outer or fifth digit of the manus and pes in all Ungulata Paridigitata (and, in
fact, in all existing Mammalia) is supported by the same carpal bone as the fourth.
There is not a single exception to this rule ; and in all Mammalia, wherever the fourth
and fifth digits of the manus and pes are present, they are always supported by one bone
— the unciform in the manus, and its homologue (the cuboid) in the pes.

Seeing that the full number of digits in Mammalia is five, while the number of carpal
and tarsal bones which give support to them never exceeds four — three inner digits
being supported each by a separate bone, while the two outer are always supported by a
single carpal and tarsal bone — the question may arise, is this relation primitive for all
Mammalia % or is it a result of coalescence of two outer carpal and tarsal bones into one 1
As we have not the slightest notion of the skeleton of the first mammals, nor even of the
geological time when they made their appearance, the solution of this question is not
to be expected now. Turning to some of the Amphibia and Keptilia with completely
developed digits, we generally, or at least very often, find, as has been proved by the
beautiful researches of Gegenbaue, that each of their digits in the manus and pes is
supported by a special carpal and tarsal bone, there being five distinct bones in the
second row in the manus and pes. Now it is very possible that the first progenitor of
the class Mammalia had also five carpal and tarsal bones in the second row, or, what is
more probable., seeing that there is not a single instance of these two outer bones being
found separate in any mammal, living or fossil, that their coalescence was effected
before the evolution of the first truly mammalian type, and that they passed into this
class already coalesced.

I have already mentioned that, beginning with the mammals of the oldest known
Eocene deposits, the oldest mammals of whose skeleton we are able to form a tolerably
clear idea (the oldest Mammalian remains, though proving the existence of the class as

* This relation is very characteristic, not only as regards Ungulata, but nearly all Mammalia.

OSTEOLOGY OP THE HTOPOTAMID^l. 69

far back as the Jurassic and even the Triassic period, have not yet furnished the slightest
clue as to their skeleton), we meet always with Ungulates which belong unmistakably
to the one or the other series ; a fossil Ungulate may be at once said to belong either
to the Paridigitate or Imparidigitate division ; we have not a single form in which the
characters of the two groups can be said to be mingled together, or which could be
reputed to be the progenitor of both these great divisions of the Ungulata. Some of
the early Eocene Paridigitata and Imparidigitata seem undoubtedly to be more nearly
related, and have more common characters, than the greatly reduced forms we meet with
in the recent epoch ; but this shows only, if we imagine these two series to be diverging
lines meeting in some old Cretaceous Ungulate, that they occupy certain positions on
these two different lines nearer to the point of divergence, but still so far from it that
all the intermediate links had time to be completely destroyed before the Eocene
period. We may imagine that in the Cretaceous epoch there existed an Ungulate
form with a very complete skeleton and five digits to the manus and pes ; from this
common form the divergence of Paridigitata and Imparidigitata may have been effected
in such a way that, with the commencement of the reduction of the skeleton of this
typical primary Ungulate, the chief development fell in one case on the middle (3rd)
digit of the manus and pes, the laterals becoming more or less reduced and arranged
symmetrically on both sides of this central digit of the manus and pes, so as to originate
the series of Imparidigitata ; while in the other case the chief development fell on
the two contiguous middle digits of the manus and pes (the 3rd and 4th), giving rise
to the series of Paridigitata. But as the overdevelopment of a single middle digit, to
such an extent as to support the body effectually, is a task much more difficult to the
organism than the development of the two middle digits to a comparatively less extent,
the reduction of the Paridigitate skeleton proceeded at a much quicker rate than that
of the Imparidigitate ; and while we meet with many animals of the Paridigitate
series having only two digits in the Eocene, the Imparidigitata have always three.
In the Miocene epoch, the two middle reduced digits of Paridigitata have already coal-
esced to form a single digit, the cannonbone of Ruminants ; but the most reduced member
of the Imparidigitata, the AnchitheriiiTn, still walked on three toes, though the laterals
began to be greatly reduced ; and it is only in the later Miocene and the Pliocene periods
that, with the appearance of the Hipparion and Horse, the skeleton of the Imparidigitata
attained as great a reduction as the skeleton of Paridigitata did in the earliest Miocene,
Qelocus (Aymard) being the first Paridigitate with a complete cannonbone in the adult.
In the skeleton of this Ungulate progenitor, before the divergence of the two series was
effected, we may imagine a pentadactyle manus and pes constructed in the way given
in the following scheme.

l 2

DR. W. KOWALEVSKY ON THE

Scheme of the Ungulate Manus and Pes.

Manus.

x 6S#

Unciform.

Magnum.

Trapezoid.

Trapezium.

Cuboid.

Cuneif.
3rd.

Cuneif.
2nd.

Cuneif.

XSl/.

Metacarpal 5th.

Metacarpal 4th.

CT*

P
o
P

ert-
P
O
P

•

CD
e+-
P
O
P
H*
h3

P
»— <

M
GO

•

£L

far*
•

CD
P

ert-
P

P
I— <

•

P
Hi

c+-

e+-

et-

•

The two outer digits of the manus and pes are supported by one bone — the unciform
in the manus, the cuboid in the pes ; the three following digits in the manus were sup-
ported, the third by the os magnum, the second by the trapezoid, the first by the
trapezium ; and in the pes each of the inner digits was supported by its corresponding
cuneiform. The reduction of the manus and pes commenced at the first or inner toe;
and we have not a single living or extinct Ungulate presenting this first digit developed ;
nay, more, I know of no positive case where even a rudiment of the first toe is present.
All the bones described as such rudiments have turned out, on close inspection, to be
the trapezium or the first cuneiform* mistaken for the rudiment of the first digit. This
is constantly the case in regard to the rudiments described in the ' Ossemens Fossiles.'

Fortunately, even in the recent period, there still exists an Ungulate upon which we
may well study the structure of the typical Paridigitate manus and pes. This is the
Hippopotamus : no trace of the first digit is left in the fore or hind limb; but, owing to
the very complete development of the remaining four digits, they must have retained
their typical relations to the bones of the carpus and tarsus. Looking at the fore foot
of the Hippopotamus (Plate XXXVII. fig. 1) we perceive that the two middle digits
are mutually interlocked in the manner described above as common to all Paridigitata.
The two outer digits, the fourth and fifth, are supported by the unciform ; the third is
supported by the os magnum, and gives besides, at its ulnar margin, a projecting beak,
by means of which it hangs on to the unciform ; the second digit goes a little higher than
its neighbours, is supported by the trapezoid, and touches a small facet on the radial
side of the os magnum ; the first digit is entirely reduced, and its proper carpal bone,
the trapezium, assists to support the second digit.

What do we see in the hind foot] The cuboid (Plate XXXVII. fig. 9, cb),
which is the homologue of the unciform, supports the fourth and fifth digits ; the inter-
locking of the two middle metatarsals is effected in the usual way, the fourth giving a
projection which enters into a concavity of the third; the third cuneiform (c 3 ) (the
homologue of the os magnum) supports the third metatarsal ; the second metatarsal is

* If this absence of even a rudiment of the first digit is really constant among fossil and living TJngulata, it
may possibly be supposed that at the branching off of the Ungulate division this first digit was already reduced
and even entirely lost.

OSTEOLOGY OP THE HYOPOTAM1DJS. 71

supported by the second cuneiform (<? 2 ), and its tibial margin touches the third cuneiform
(just as the homologous parts on the fore foot do) ; the first digit is entirely lost, and its
typical tarsal bone, the first cuneiform, aids in supporting the second digit (Plate XXXVII.
fig. 10, c x ).

This should be the place to consider the structure of the feet in the Suina ; but, as in
the recent members of this family, owing to the overdevelopment of the middle digits,
the typical relation between the bones is a little changed, I will describe it after having
treated those Paridigitata which preserve this typical relation unchanged.

Of the fossil Paridigitata whose skeleton is known to us with any thing amounting to
completeness, we have only the Anoplotherium and Xiphodon ; and we shall see by-
and-by with what wonderful persistence the typical relations are adhered to in the
skeletons of these animals, notwithstanding the great reduction of the number of their
metacarpals and metatarsals.

As the Anoplotherium tridactylum is certainly a less-reduced form than the species from
the gypsum, and as, except a sketch in Gervais's ' Paleontologie Fran§aise,' there are no
good drawings of its extremities, I represent them on Plate XXXVII. figs. 2, 11, from the
original specimens of Bravard now in the British Museum. Though the number of digits
on the fore and hind limbs is odd, it is nevertheless a very typical Paridigitate, and
differs from the Paris Anoplotherium only in this respect, that the second digit, which is
represented only by a small rudiment in the animal from Montmartre, is developed to
a complete (though short) digit in the Anoplotherium from the lignites of Vaucluse.
The manus of the Anoplotherium is so well described by Cuvibr that I will not enter into
any details, and only point out its chief peculiarities. As seen in Plate XXXVII. fig. 2,
this manus is entirely true to the general type : the interlocking of the two middle meta-
carpals is effected in the same way as described above ; the fourth metacarpal is supported
by the unciform ; the third hangs to the os magnum, and sends a prolongation of its
ulnar margin to meet the unciform; the second, though too short to touch the ground,
is complete and supported by the trapezoid, and it sends a projection to articulate
with the os magnum ; the trapezium aids in supporting the short second digit. At the
outer margin of the foot there exists a rudiment of the fifth digit, leaning against the
unciform and the ulnar margin of the fourth metacarpal.

The pes of Anoplotherium tridactylum (Plate XXXVII. fig. 11) shows us the same
typical persistence in the relation of the bones. The interlocking of the two middle meta-
tarsals takes place, as usual, by a process from the fourth fitting into an excavation of the
third metatarsal ; this last is supported by the third cuneiform (c? 3 ). As the second digit
is developed, the corresponding bone, the second cuneiform (<? 2 ), very small or nearly obso-
lete in the Paris Anoplotherium, is largely developed in the A. tridactylum, and supports
the second digit ; the first digit is absent, and its tarsal bone, the first cuneiform (<?,),
aids in supporting the second digit.

As the chief difference between the Anoplotherium from the gypsum of Paris and
the A. tridactylum lies in the development of the second digit and its corresponding

72 DE. W. KOWALEVSKY ON THE

carpal and tarsal bones, I have represented the distal surface of the navicular (to which
the cuneiforms are attached) of the tridactyle form (fig. 19) and of the Paris Anoplo-
therium (fig. 17); the navicular (fig. 18) is from the same locality as the tridactyle
one ; and though undoubtedly belonging to a didactyle Anoplotheriuni, the intermediate
facet for the second cuneiform still presents some development, while in the Paris spe-
cimen it is exceedingly small and is even absent in some specimens ; indeed it seems that
the second cuneiform, being very small, did not always touch the navicular ; and this led
Cuviee into the erroneous belief that this bone was altogether wanting*.

The manus and pes of Xiphodon (Plate XXXVII. fig. 3) present exactly the same
peculiarities : though reduced to only two metacarpals and metatarsals, these two are
in no way adapted more completely to the distal surface of the carpus, and are sup-
ported only by their typical carpal and tarsal bones. The persistence of typical rela-
tions is so great that even the rudiment of the second metacarpal, though a mere bony
nodule, persists not only in its articulation with the trapezoid, but even with the radial
facet of the os magnum (as seen in Plate XXXVII. fig. 3, u., m, t), as truly as in the
case of the complete second metacarpal in the four-toed foot of a Hippopotamus.

In the pes, as far as I am able to see by the figures of Cuviee and De Blainville, the
relation between the metatarsal and tarsal bones is quite such as we laid it down in
our general scheme. There are three distinct cuneiforms ; the third metatarsal is sup-
ported entirely by the third cuneiform ; and though there is on]y a small rudiment of
the second metatarsal, nevertheless this rudiment has not surrendered its typical articula-
tion, and retains the whole of the second cuneiform for itself; the first cuneiform is arti-
culated upwards with the scaphoid, touches the posterior part of the second cuneiform,
and is articulated lower down with the rudiment of the second metatarsal, presenting in t
this way entire agreement with the typical structure of a Paridigitate tarsus f.

Now, if we turn to Cuviee (Oss. Foss. v. 4to ed. p. 181), we shall see that he noticed
the three cuneiforms in Xiphodon, calling only the first (marked e) " osselet sur-
numeraire ;" he did not notice the true rudiment of the second metatarsal (De Blain-
ville is entirely wrong in saying that there were only two cuneiforms, Anopl. p. 50) ;
he did not describe or mention the rudimentary second and fifth metatarsal, though he
very correctly noticed the corresponding bones in the metacarpus.

After this brief notice of the structure of the metapodium in the fossil Paridigitata,
we may proceed with the description of the same parts in Diplopus and Hyopotamus ;
but I wish, before doing so, to take a short survey of the same part of the ske-
leton in the remaining living representatives of the Paridigitate series, the Suidae and
Euminantia.

* Passing through Paris, I tried to settle this question ; and in fact there exists in the < Galerie de Pale-
ontologie ' a nearly complete foot with the three cuneiforms. It is in a block of gypsum, and placed high on
the top of the wall-cases in the gallery.

t I find on the posterior part of the distal articular surface of the cuboid of Xvphodon a small facet clearly
destined for the rudimental fifth metatarsal.

OSTEOLOQY OF THE HYOPOTAMID2E. 73

The general conclusion to be drawn from our survey of the fossil forms, is that
they exhibit an extreme uniformity in the structure of their metapodium, and that these
uniform and typical relations between the metapodium and the bones of the carpus and
tarsus are remarkably constant, and, as regards the fossil forms, unpliant and rigid. Even
with the utmost reduction of the metapodium in Anoplotherium and Xiphodon, the
typical relation between the two remaining metacarpals and metatarsals remains
just the same as it probably was in their tetra- or pentadactyle ancestors ; so long as
even a rudiment of a metacarpal or metatarsal remains, it holds just the same re-
lation to the supporting bones as if it were complete. We meet with no distinct
adaptation by means of which the median metacarpals and metatarsals which are left
after the dropping off of the laterals, enter into a more complete articulation with
the remaining bones of the carpus and tarsus. Considering, for instance, the two slender
separate metacarpals and metatarsals of Xiphodon, we must confess that a foot so
badly adapted for the use of a swift animal is rarely to be met with The middle
digits, being unankylosed, are liable to be broken separately by a much less exertion
of force than it would require to break the two coalesced middle digits of a Ruminant ;
besides, seeing in what manner the whole weight of the body is transmitted to the two
middle digits, we shall find that it is not effected in such a way as to ensure the
most complete and stable equilibrium. To effect this we might expect that the
proximal ends of the two remaining metacarpals (Plate XXXVII. fig. 3) would be
enlarged to such a degree as to underlie the whole distal breadth of the carpus ; in this
case the weight of the body would be transmitted much more equably and effectually to
the two middle digits of the metacarpus. However, we see nothing of the kind ; the
transmission of the weight of the body is effected only by the two bones of the second
row of the carpus ; and the two useless rudiments remaining on both sides, and occupying
the whole trapezoid and a large facet of the unciform, diminish by so much the stability
of the foot, in comparison with an arrangement in which the facets occupied uselessly
by them should be taken by the functional middle digits. We meet with exactly the
same relation in the pes ; so that it will be needless to recapitulate in reference to it all we
have said in reference to the manus. The same may be, to a great extent, said of the
Anoplotheriuni, though the digits which remain are so stout and short that the want of
stability of the foot is not so clearly shown by this form as by its slender congener.

On the whole we may, with sufficient probability, say that, while in these two genera
the reduction of the manus and pes was going on and the lateral digits aborted,
the remaining middle digits did not adapt themselves as fully -as could be imagined* to
altered circumstances of life and to a different distribution of weight of the body; they
remained too true to ancestral traditions ; there was no pliancy in their organization
which, by adapting them more fully to altered conditions of life, would have enabled
them to carry on successfully the struggle for existence with other competing genera.

*■ And as we actually see in other genera which outlived Anoplotherium and Xvphodon.

74 DE. W, KOWALEVSKY ON THE

And these last, as we shall presently see, being better adapted to altered conditions of
life, got the upper hand, multiplied largely in specific and generic forms, and peopled the
earth with their successors, while Anoplotheriurn and Xiphodon died away without leaving
any*. In this inflexibility and rigidity of organization, in this inability to alter it as
completely as the competing genera were able to do, lay perhaps one of the causes of the *
extinction of some genera and their replacement by others. All that I attempt here is
to show in what peculiarities of structure this rigidity of some genera and pliancy of
others consisted. I do not wish to put this as the only cause, but as one of the many
still unknown causes which led to the extinction of so many animals that preceded the
present population of the earth.

I have adduced here only such cases as are known and described ; but having carefully
examined large collections of bones from the Eocene and Miocene deposits, I have been
struck with the recurrence of similar facts. Trying to reconstruct the extremities of some
Paridigitata of the Lower Eocene from Mauremont and Egerkingen, and of Miocene forms
from Eochette, I could distinctly perceive that all genera which have left no direct succes-
sors, and which are entirely extinct, present the same rigidity and persistence of the
typical relations ; on the contrary, those which have representatives in the living creation,
their direct successors, exhibit much more pliancy and much better adaptation to altered
circumstances of locomotion, along with the reduction in the number of digits.

Before proceeding to consider the two remaining groups of Paridigitata, the
Euminantia and the Suina, we may draw the attention of the reader to the fact
that these two groups of Paridigitata are the only onesf which now people the
earth ; there is no greater diversity than this ; and every living Paridigitate (if we
except the Hippopotamus) is always clearly a ruminant (including the Tragulidse and
Camelidse) or a Sus. How striking is this poorness of different types if we compare it
with the rich and diversified forms presented by the recent Carnivora or Eodentia. The
extreme diversity of generic forms and specific modifications, coupled with the enormous
range of distribution of living Paridigitata, produces a false impression of the diver-
sity presented by this order ; but in reality there is no such diversity, and all the extremely
rich assemblage of Paridigitate Ungulates that people the earth in our time are only the
result of the modification of two typical forms, the Suina and the Euminantia. The
latter term is a very objectionable one, as the faculty of rumination has nothing to do
with the skeleton, and in reality it would be no wonder if some of the Imparidigitates
possessed the same faculty. As the teeth have so great a value in systematic zoology,
it would be perhaps more advantageous to distinguish all Paridigitata into those
which have tubercular and those which have crescentic teeth. To the first division will

* The reduction of the limbs in Anoplotherium and Xiphodon is so great that I regard them only as the last
representatives of dying- out branches that did not leave any direct descendants.

f If we except Hippopotamus — this last remnant of a Paridigitate series once rich in generic forms, but
which is now reduced to only two distinct groups,

OSTEOLOGY OF THE HYOPOTAMID2E. 75

belong all the existing Suina and the 'Hippopotamus ; to the second, the remaining
Paridigitata, which all possess, more or less completely, the faculty of rumination
coupled with the absence of incisors (Camel ]) in the premaxillaries. Such a division
of Paridigitata would allow a place in the zoological scale to the extinct forms which
were distinguished by the non-confluent metapodium, and the presence of incisors in the
premaxillaries, and very probably did not ruminate.

We shall now proceed to cast a comparative glance at the structure of the feet in
both these divisions as they exist in our own time, and endeavour to discover if their
structure does not present some characters which show that they are better adapted for
new circumstances of life than were their Eocene and Miocene predecessors; and that
to this better adaptation may be, in part, ascribed the victory they obtained in the battle
of life, and their spreading over all the surface of the globe.

As the Paridigitata with tuberculated teeth are represented, in the recent period,
only by the Suina, and those with crescentic teeth only by the Euminantia, we shall
have to confine ourselves to these two orders ; in the latter we shall particularly call
the attention of the reader to the Tragulina, as the less reduced members of this
family, and therefore more likely to furnish us with typical characters.

The true Suina have four complete digits in their manus and pes, but only the two
middle ones are subservient to the purpose of locomotion ; the laterals are always so
reduced that they do not regularly touch the ground, or only do so on muddy soil, when
the foot sinks deeply into the earth. According to the general rule laid down for all
the Paridigitata, the interlocking of the two middle (third and fourth) metacarpals in
the manus is effected as usual ; the fourth digit is supported by the unciform, while its
radial upper margin is fitted into an excavation on the ulnar side of the third metacarpal,
which, by means of an ulnar process of its upper margin, articulates with the unciform,
while its proximal surface is supported by the os magnum (see figs, in Cuvier, ' Oss. Foss.
Atlas,' and De Blainville, ' Osteographie, SusJ also our fig. 4, Plate XXXVII.). How-
ever, in examining more attentively this proximal surface of the third metacarpal of
the Hog, we remark something quite new, and not met with in most of the fossil
Paridigitata. Owing to the over-development of the middle digits, the radial side of
the third metacarpal (Plate XXXVII. fig. 4) has spread inwards and pushed the
second metacarpal away from its typical articulation with the os magnum; nay even
more, this second metacarpal has yielded one half of the surface of its carpal bone
to the encroachment of the third digit; this last, besides the magnum, occupies
now one half of the trapezoid — a new fact in the history of the Paridigitate foot
that had important consequences. To show the reader more clearly that this modifica-
tion is a recent one, we turn to the fossil Suidse. Unhappily, our knowledge of their
skeleton is very imperfect; and while genus after genus (not to speak of species) of the
fossil Suidse have been created merely on dental, often very slight and unimportant,
characters, the study of their skeleton has been much neglected. As far as I am aware,

MDCCCLXXIII. M

76 DE. W. KOWALEVSKY ON THE

not a single bone of their skeleton was figured or discussed comparatively until the
appearance of Professor Gaudry's work on the Fossils of Pikermi*.

However, I was so fortunate as to see many bones of the skeleton of the Miocene
Suina, distributed into divers genera, in the collections of Paris and London ; and on
comparing the third metacarpal of the Palwochcerusf (Plate XXXVII. fig. 6) from
Auvergne, and the Cheer omorus, Lrt, from Sansans, with that of the recent Suidae, I
found that this broadening of the third metacarpal did not exist in Palmochoerus nor in
the Cheer omonts, the second digit taking for its support the entire surface of the trape-
zoid. Plate XXXVII. fig. 6 represents a third metacarpal of Palwochoerus ; and
by comparing it with fig. 4 we immediately perceive the difference. This broadening
of the metacarpals, with the purpose of taking the whole of the distal surface of the
second row of bones of the carpus, took place in the Suina only after the middle Miocene
epoch. But, among the recent Hogs, we have one more advanced form which, in this
character, stands to the other Suidse nearly in the same relation as the Suidse stand to
the Paleeochoeridse ; this is Dicotyles. The structure of its fore limb agrees entirely with
that of Sus, only the lateral metacarpals are reduced a step further ; and looking at
fig. 5, Plate XXXVII., we shall see that the enlarged third metacarpal has taken, not
the half, as in Sus, but the whole of the distal surface of the trapezoid, the second
metacarpal being pushed entirely away from its typical carpal bone. Moreover, this facet
for the trapezoid (as in fig. 4) is transformed from an oblique into a horizontal one,
thus giving better support to the third digit ; while the corresponding distal surface of
the trapezoid instead of having a spear-shaped form, as in the Hog (fig. 4, #), is quite
flat in Dicotyles (fig. 5, t). The trapezium (fig. 5, tz) is greatly reduced, has no distinct
articular facet on the trapezoid, and is disappearing altogether, without coalescence with
the trapezoid. Thus, then, we find that, in the Miocene Pigs, the whole trapezoid is,
according to the typical relations, taken by the second digit ; in the recent Hog only
half of its distal surface is left for the second digit, and none at all in Dicotyles.

On the outer side of the manus, the relations are much simpler, as the fourth and
fifth metacarpal are supported, in all Ungulata, by a single bone, the unciform ; there-
fore, by the gradual broadening of the fourth metacarpal, the outer or fifth digit is
pushed to the outer side of the distal surface of the unciform, until, in the Hog and in
Dicotyles, it occupies only a small lateral and outer facet on this bone, in such a way
that the fifth digit has practically no upper carpal facet, but is suspended laterally to

* Gebvais, 6 Paleontol. Erancaise,' has figured a third metacarpal, which is very interesting, as it shows none of
the broadening of the radial margin so characteristic of Sus. It is from the right side, while all figured by me
are left.

f In its dental characters, as well as in its skeleton, Pakeochcerus stands so near to the Hog that even
their generic distinction might be questioned. The Choeromorus, being also a true Hog, stands further from
the recent Suidae ; one of the very curious characters of this remarkable genus is the central ridge of the distal
extremities of the metacarpals and metatarsals, which, instead of encircling the whole extremity, is limited to
its back part only ; the first phalanges are modified accordingly.

OSTEOLOGY OF THE HYOPOTAMIME. 77

the outer side of the unciform ; and this lateral position is more pronounced in Dicotyles
than in the Hog, in consequence of the greater reduction of the former.

Moreover, if we examine attentively the proximal surfaces of the two middle meta-
carpals in Dicotyles, we shall see that they are much more joined together than in the
common Hog or any other Sus ; their inner flat sides are so closely united that it will
require only a little step further to make them coalesce; and as the os mapum and
Jpezoid L now bo«L Sti „ g on one meUcarpa! they eannot remain .ong' distinct,
but must coalesce : if this should occur, we should have a structure nearly analogous
to, and hardly to be distinguished from, that of a typical ruminant.

If we turn now to the structure of the pes in the Paridigitata with tuberculated teeth,
or Suina, we shall meet with precisely the same phenomena ; and the homologous bones
of the manus and pes undergo a strikingly similar course of variation.

The two middle digits of the pes interlock, in the usual way, by a process from the
fourth metatarsal, which fits into an excavation of the third. The cuboid supports the
two outer digits, the fifth and fourth ; the third digit is supported by the third cunei-
form. If we compare the third metatarsal of the recent Hog (Plate XXXVII.
fig. 12, in.) with the same metatarsal of Palwochcerus, fig. 12', and of Cheer vmofus, we
perceive just the same difference as in the homologous bones of the manus. While in
Palwochoerus the third metatarsal is confined entirely to the third cuneiform # , leaving the
second cuneiform for the support of the second metatarsal, in the Hog (fig. 12, in. c 39 c 2 )
this third metatarsal is greatly enlarged ; it has pushed the second digit away,
and encroached on nearly the whole of the distal surface of the second cuneiform,
leaving only a very narrow facet of this bone for the second metatarsal (fig. 12, n. c 2 ) 9
which is now chiefly supported from behind by the wedge-shaped first cuneiform.

On the outer side of the pes, the enlargement of the fourth metatarsal has taken the
greater part of the distal surface of the cuboid, the fifth metatarsal being pushed very
much backwards, and being supported partly by a small facet on the distal surface of the
cuboid, and partly by a posterior prolongation to the beak-like downward process of the
same bone. The pes of the Dicotyles shows us the same disposition in an exaggerated
form; the inner side of the third metatarsal (Plate XXXVII. fig. 13, in.) has completely
occupied the second cuneiform, so that the second digit is supported only from behind by
the first cuneiform. On the outer side, the fifth digit is completely lost ; or, if a rudiment
of the fifth metatarsal remains, it is generally a flat elongated bone attached to the outer
side of the fourth metatarsal, and having no articular surface on the cuboid, which is
taken up by the over-developed fourth metatarsal. Besides, the two middle digits
have coalesced in the whole upper half of their length, simulating the cannonbone
of a Ruminant. The navicular and cuboid are still separated, but are very firmly
pressed together; and now, both these bones, having the coalesced metatarsal below
them, cannot have much separate movement, and their coalescence, as well as that of the

* As the hind limbs are always more reduced than the fore, it seems that, even in Palwochoerus } the third
digit encroached in a perceptible way on the second cuneiform.

M 2

78 DE. W. KOWALEVSKY ON THE

second and third cuneiform, is only a question of time, the modification going on unin-
terruptedly in the direction of the greater reduction of the limb-bones.

Can the general tendency of this steady reduction be doubtful] Is not the fact
eloquent enough, that, proceeding from the middle Miocene times until the recent
period, we meet with a whole series of Suina in which the skeleton is gradually more and
more reduced, until it culminates in the Post-tertiary time in Dicotyles, a form very
analogous, in the structure of its limbs, to the Ruminants, whose middle digits are quite
ready to coalesce into a cannonbone, and the laterals to drop off \ Indeed, this has begun
in the posterior limb, in which reduction is always in advance, and on its outer side, which
is generally more reduced than the inner. If any inference from one series of pheno-
mena is allowed to be applied to another, then, inasmuch as the Paridigitata with cres-
centic teeth, or the recent Euminantia, proceeding from tetra- or even pentadactyle
forms, arrived in the Miocene period at didactyle forms, in which the coalescence of
the two middle digits simulates monodactylity, we have a full right to infer, seeing
the parallelism of these two groups, that the Paridigitata with tubercular teeth
have followed just the same line of reduction. And if nature should be allowed to
follow its course, or if man had made his appearance only in the Post-quaternary instead
of the Post-tertiary period, he would no doubt have found only two groups of Pari-
digitates remaining, one with crescentic, the other with tubercular teeth, but both having
a cannonbone in their fore and hind limbs, and no lateral digits. These two groups
of Paridigitates undergo exactly parallel modifications in the course of time, as far as
their limbs are concerned— only in the group with crescentic molars these modifications
have gone on much more rapidly than in the parallel group with tubercular teeth. The
cause of this greater rapidity lay very probably in the more specialized, instead of an
omnivorous, diet, and was perhaps influenced by the commencing faculty of rumination,
which gave them an enormous advantage over the other group, by allowing them to
store food in their paunch in the most favourable, or least dangerous, part of the day,
and chew it afterwards when retiring to rest.

If we turn now to the Paridigitates with crescentic teeth, represented in our times
only by the living Euminantia, we meet in the typical (which in this case means the
most reduced) genera both middle metacarpals and metatarsals coalesced so as to simu-
late a monodactyle foot, forming the so-called cannonbone. The rudiments of the
lateral digits are mostly lost (in Bovidce and Antilopidce) or retained only as small
filaments of bone, having no articulation with the carpal or tarsal bones, but merely
pressed to the outer and inner sides of the two middle digits. The trapezium is
entirely lost ; the trapezoid is confluent with the magnum *. In the tarsus, the navicular
is confluent with the cuboid, and the second cuneiform with the third ; the first remains
nearly always separate, and in case of confluence with the coalesced second and third
cuneiforms, as in the Giraffe, the division is clearly seen. This is the general structure
of the foot in the typical Euminantia, or the most reduced Paridigitata with crescentic

* Except in Camelidce.

OSTEOLOGY OF THE HYOPOTAMIME. 79

teeth ; but fortunately we have still a living form which stands to the typical Kumi-
nantia nearly in the same relation as our Dieotyles would stand to the Post-quaternary
Suinae with a cannonbone. The parallel is really complete, with the exception of
some small points*. This living form is the Hyomoschus aquaticus-, hardly distinguish-
able from its fossil congener of the Middle Miocene. If we examine the manus of this
animal (Plate XXXVII. fig. 8), from the inner side, we shall meet with characters
common to all Paridigitata. The interlocking of the two middle metacarpals is
effected in the usual way ; the inner, or radial, margin of the third is enlarged in the
same way as we saw it in Dieotyles — only the two carpal bones which support this
enlarged third metacarpal, the magnum and trapezoid, are already confluent (fig. 8>
m & td) ; the reduced, but still complete, second digit has a small facet on the back part
of the coalesced trapezoideo-magnum. On the outer side of the manus, the large fourth
metacarpal is supported by the unciform, and the distal surface of this bone gives also
a small facet to the reduced and thin, but still complete, metacarpal of the fifth digit.

The pes of Hyomoschus will show us something similar to what we have seen in the
manus. As seen in fig. 14, Plate XXXVII., the inner side of the third metatarsal is
enlarged, and has taken the whole of the second cuneiform, the second metatarsal being
supported entirely by the first cuneiform, which is distinct, while both the others have
coalesced mutually and with the navicular (c s +c 2 +n, fig. 14).

On the outer side of the pes we find that the large fourth metatarsal has taken nearly
the whole distal surface of the cuboid ; however, it leaves a very small facet for the
articulation of the complete fifth digit. The length of the lateral metacarpals and
metatarsals nearly equals that of the middle ones, though, owing to their thinness and
the want of direct firm support from the carpals and tarsals, they are, as it seems, not
subservient to locomotion. The middle metatarsals are confluent in the adult.

In the Tragulidse (Plate XXXVII. figs. 7 & 15) the middle metacarpals and meta-
tarsals are distinct in the young, even after their complete ossification; in this state we
may ascertain that their mutual interlocking is effected as in all other Paridigitata.
The inner margins of the third metacarpal and metatarsal are enlarged even more than
in Hyomoschus, and their relation to the trapezoid and second cuneiform is altogether
the same. The lateral digits persist during the whole of life as filiform bones on both
sides of the middle cannonbone. Although of the same length as the cannonbone, they
are useless for locomotive purposes, owing to their extreme thinness. At last, in the
typical Euminantia, the two middle metacarpals and metatarsals coalesce into the cannon-
bone during the process of ossification. All particulars we have remarked in the Tragulinse

* The chief are these : — the trapezoid is confluent with the magnum, while it is only preparing to become so
in the anomalous Dieotyles*, the navicular of the pes is confluent with the coalesced second and third cuneiforms
(as in all the Tragulidse) ; the fifth digit of the pes is completely developed, not lost, as in Dieotyles. The
articular ridge of the distal end of the metacarpals and metatarsals is confined only to the palmar side, and
the first phalanges are modified accordingly. If we remember that some of the ancient Pigs, as Ghceromorus>
show just the same smoothness of the distal ends of the metapodium, the parallelism of both groups of
Paridigitata appears to hold good, even in the slighter details.

80 BE. W. KOWALEVSKX ON THE

are exaggerated in the true Ruminants. The trapezoid and the second cuneiform
are always, even in the cartilaginous state, confluent with the magnum and third cunei-
form, and can be detected only as distinct points of ossification. The rudiments of the
lateral digits are generally preserved as slender elongated bones in the metacarpus^
especially in Deer, much more seldom in the metatarsus. In the fossil Euminantia
from Auvergne these lateral rudiments are present, as a rule, on both the fore and hind
limb (Plate XXXVII. fig. 16, il). The rudiment of the fifth metatarsal is usually free,
and has even a small facet on the cuboid; while the second metatarsal, being jammed
in between the inner enlargement of the third metatarsal and the posterior beak-like
prolongation of this digit, has generally coalesced; but its proximal extremity is mostly
free, and articulated with the first cuneiform (and this is sometimes to be seen even in
living Euminantia).

The upper and posterior beak-like prolongations of the middle metatarsals of Pari-
digitata constitute a very characteristic feature of these bones ; they grow larger and
larger with the reduction of the pes, and are firmly pressed together by a special process
of the cuboid and the first cuneiform in those genera in which locomotion is almost
entirely performed by the two middle digits. Finally they coalesce, and the confluence
of the metatarsals seems to proceed from these processes downwards, as is to be seen at
a certain age in the metatarsus of Dicotyles.

Metacarpus and Metatarsus of Hyopotamus and Diplopus.

Having thus discussed at some length the shape and mutual connexions of the bones
which compose the fore and hind limbs in the chief fossil and living Paridigitata,
our task is made much easier in reference to IHplopus and Hyopotamus ; and without
dwelling too long on the description of very minute particulars of these bones, which to
a certain extent are visible in the Plates, I shall merely state the chief features they
present to the observer, and the points of agreement or difference with the corresponding
bones of allied genera.

I begin -with the metacarpals of Hyopotamus, or the tetradactyle form, and will
afterwards pass to the didactyle Diplopus. The restoration of the manus of Hyopotamus
from Puy is given in Plate XXXVII. fig. 20, J nat. size, and the upper part of the
same manus, fig. 5, Plate XXXVIII. The separate bones belonged to different indi-
viduals, and were not found in connexion. The fourth metacarpal especially is defective ;
but it is the only specimen of this digit I could find in the collections from Puy ; it comes
from a very young individual, and is therefore too small for the adult unciform by
which it is supported. I have seen the same bones from Hempstead, but somewhat rolled.

The interlocking of the two middle metacarpals is effected in the usual way, the
radial margin of the fourth being thrust under the ulnar prolongation of the upper
margin of the third (Plate XXXVIII. fig. 5). This ulnar prolongation of the third
metacarpal going to meet the unciform is very oblique in the Hyopotamus ; its axis
forming, approximately, an angle of 45° with a horizontal line. In the Suidee, as well

OSTEOLOGY OF THE HYOPOTAMIIUE. 81

as in the corresponding bone of Diplopus (Plate XXXVIII. fig. 6, in.), this projection is
much more inclined (perhaps 30° with the horizon). The reason of this difference is given
by the unciform, as the radial inferior truncature of this bone (Plate XXXVIII. fig. 5, u),
to which the projection of the third metacarpal is articulated in Eyopotamus, becomes
nearly vertical, and helps to constitute the inner radial wall of the same bone in Diplopus
(Plate XXXV1I1. ng.-8).

The proximal surface of the third metacarpal (fig. 5, in.) has an elongated, somewhat
triangular, fiat facet for the os magnum ; this upper surface is sloping inwards, so as to
allow the second metacarpal to lean laterally on this margin, and reach its facet on the
os magnum, as seen in fig. 5. In Diplopus, on the contrary (fig. 6, in.), this upper
radial margin of the proximal surface is produced upwards (as in Palceoclicerus, Plate
XXXVII. fig. 6) ; and therefore I think that, in the didactyle form, the rudiment of the
second digit could not reach as high as the os magnum,

The ulnar side of the third metacarpal has an anterior reniform and a posterior oval
facet ; both articulate with corresponding facets on the radial side of the fourth digit.
In the didactyle Diplopus, this anterior reniform facet is excavated into a deep hole for
a corresponding large projection of the fourth metacarpal, as seen in Plate XXXVIII.
fig. 6 : thus the interlocking of the two middle digits is much closer in the didactyle genus ;
and this is quite natural, as the foot, having no lateral digits, required a firmer structure.

The radial side of the third metacarpal in Eyopotamus has on its anterior part a tole-
rably long (7millims.) facet, which is uninterruptedly united with the somewhat sloping
radial margin of the proximal surface. . This facet is destined for the second metacarpal,
which has a corresponding surface on its inner (ulnar) side ; the bone on this side is flattened
by the pressure of the complete lateral digit, as may be seen in the sections of the four
metacarpals given below the fig. 5. If we compare this part of the third, meta-
carpal with the corresponding region in Diplopus (fig. 6, iil), we shall see a great
difference. In this last genus, as there is no lateral second metacarpal, but only a rudi-
ment of it, we find a deep depression, with small longitudinal facets, where the rudiment
(seemingly a nodular bone, as in Anoplotherium) adhered to the third metacarpal. We
meet with an exactly similar hollow on the inner (radial) side of the third metacarpal in
Xiphodon. On the upper and front part of this third metacarpal is a rough tuberosity for
the tendon of the extensor carpi radialis muscle ; this is not so prominent in the same
bone of Diplopus Aymardi (fig. 6), perhaps owing to the younger age of the individual
to which the third metacarpal of fig. 6 belonged.

The fourth metacarpal (Plate XXXVIII. fig. 5, iv.) of Eyopotamus, though repre-
sented very badly by a proximal half of the bone belonging to a young individual
from Puy, shows, nevertheless, nearly all we require to know about it.

The proximal surface is flat, and has the shape of an isosceles triangle. On the radial
upper margin we have a thickening at the point where the fourth digit is thrust under
the ulnar prolongation of the third. The ulnar, or outer, side of the fourth metacarpal
has two distinct facets, an anterior and a posterior one, for the articulation of the outer
or fifth metacarpal.

82 DE. W. KOWALEVSKX ON THE

The general shape of both middle digits in Hyopotamus is very fiat in front, with two
tolerably sharp edges, which form the outer and inner border of each metacarpal ; the
two inner, or contiguous, edges are formed by the metacarpals being pressed againsteach
other in the axial line, while the outer edges result from the pressure of the lateral digits
on the two middle ones from without and within (see fig. 5, and section). In the
Dijolojous, as we shall see hereafter, we have only the contiguous edges, while the outer
sides of both metacarpals are rounded and smooth, there being no lateral digits. The
section of the four metacarpals given below fig. 5 may give an idea of the flatness of
these bones in the middle : it may be possible that they are somewhat flattened by
pressure ; but in the living Hippopotamus the metacarpals are perhaps relatively as flat.
The length of the third metacarpal (drawn in outline) from Puy is given by a com-
plete specimen in the collection of M. Aymard. I have also several rolled specimens
of both middle and internal lateral digits from Hempstead.

The distal ends of the two middle metacarpals (Plate XXXVII. fig. 20) are quite
smooth in front ; but on the posterior or palmar surface of each a median prominent ridge
is seen corresponding with a sulcus on the proximal extremity of the first phalanx. In the
Hog, this ridge goes round the whole distal end of the middle metacarpals. The outer
half of this distal end is a little shorter than the inner, though the difference is not, by
far, so great as in Suina ; and the first phalanges being modified accordingly to fit the dis-
tal extremity of the metacarpals, the outer half of the proximal surface of each is slightly
higher than the inner. This corresponding inequality of the distal ends of the meta-
podial bones and the proximal surface of the first phalanges is seen in many fossil genera,
but is most manifest in the recent Suina. The purpose of this arrangement seems to be
the compression of the metacarpals of the middle digits and the approximation of their
distal ends. The two middle digits diverge in treading on the ground ; but, by the same
action, their proximal extremities tend to converge ; and, owing to the special disposition
of the metacarpophalangeal articulation, they compress the two metacarpals or meta-
tarsals together. In such animals as the Hog, in which this disposition is strongly de-
veloped, and aided by a special adaptation of the tarsal, and to some extent carpal, bones
to bring together the proximal extremities of the two middle digits, these two practically
work like the cannon of a Ruminant. In Hyopotamus, however, this disposition is
only indicated, while it is developed much better in IHplopus.

The lateral digits of the Hyopotamus are very well developed, and, in fact, besides the
Hippopotamus, we know of no animal in which they are so complete and relatively large
as in Hyopotamus. It is possible that in Anthracotherium they took even a more im-
portant part in locomotion ; but the complete limbs of Anthracotherium are not fully
known at the present time.

The inner or second metacarpal (Plate XXXVII. fig. 20, & Plate XXXVIII. fig. 5, n.)
is represented in all collections I have visited only by its proximal half. The upper or
proximal extremity of the second digit presents an elongated and concave articular
surface for the distal face of the trapezoid; its ulnar margin, as seen in fig. 5, is
blunted by an oblique facet, which must have abutted against the os magnum.

OSTEOLOGY OF THE HYOPOTAMXD.E. 83

On the posterior edge of this second metacarpal is a small facet, probably for the
trapezium ; although I did not find this last bone, its existence must be inferred from a
facet on the trapezium and this posterior metacarpal facet.

The general shape of the bone is somewhat triangular ; on its flat side it is pressed
against the third metacarpal.

The fifth metacarpal. — Of this bone I have seen only somewhat less than the upper
third. As shown by this small fragment, the fifth digit was relatively well developed ;
our fragment probably comes from a larger individual than the other metacarpals.
The proximal surface has a facet which occupies its entire antero-posterior depth and
articulated with the unciform (fig. 5, v.); the inner side shows two separate facets, one
anterior and one posterior, which entirely correspond with similar facets on the outer
side of the fourth metacarpal (the Hog and Hippopotamus has only one such facet). The
outer edge of this fragment is thickened and rugose for the attachment of ligaments.

The Metacarpus of Diplopus.

The chief differences exhibited by the two middle metacarpals of the didactyle
Diplopus (Plate XXXVIII. fig. 6) have been to a certain extent already stated in the
course of the discussion of the metacarpals of Hyopotamus; something may, however,
be added. The general shape is exceedingly different, as may be seen by the sections.
The two metacarpals of the didactyle Diplopus are much more mutually symmetrical
than those of Hyopotamus ; the contiguous sides are flattened in such a way that the
two bones are pressed together on their flat surfaces, while their outer rounded outlines
sweep in and out in a nearly regular quadrant, so that the two united metacarpals
represent in section the half of a solid cylinder. These inner flattened faces are very
rough, showing the attachment of numerous ligaments that held them firmly together.
The distal extremity is turned a little outwards and broadened transversely ; its inner
half is much thicker or deeper than the outer, more so than in the recent Suina.
The distal articular ridge of the metacarpals is limited only to the posterior or palmar
side, although a faint trace of it is visible even on the anterior face of the distal
extremity of the metacarpals. The semicircular line, where the anterior surface of the
metacarpal passes into the distal articulation for the first phalanx, is only slightly
excavated, while in the metatarsals this line presents a deep crescentic concavity; this
furnishes a very good practical character for distinguishing broken distal ends of the
metacarpals from the metatarsals.

We have already mentioned the difference in the proximal surface of the third
metacarpal, by the radial edge (fig. 6, in.) being more raised, by the more horizontal
direction of the ulnar process, and by the depth of the excavation, into which is
fitted the radial projection of the fourth digit. The posterior surface of the meta-
carpal is flat, as seen in the section. The fourth metacarpal (Plate XXXVIII. fig. 6,
IV.) shows similar differences ; the proximal surface is of a rounded triangular outline
to fit the similar facet of the unciform (fig. 8', iv.) ; the projection on the radial flat

MDCCCLXXIII. N

84 DE. W. KOWALEVSKT ON THE

surface is very prominent, in order to enter into the deep pit on the ulnar side of the
third and bring about a firm interlocking of the two digits. On the outer and posterior
side, we see an excavation for the nodular rudiment of the fifth metacarpal, which, as it
seems, did not touch the unciform, at least had no facet on it. The posterior surface
is very nearly flat, and the section of the bone is a rounded triangle instead of being a
flat trapezium as in the Hyopotamus. I have seen a complete specimen of only this single
metacarpal ; fortunately this gives us the length of the metacarpus in the didactyle
genus ; it is considerably longer than the metacarpus of Hyopotamus^ as may be seen
by comparing figs. 5 and 6 of Plate XXXVIII.

As we have no means of distinguishing with complete certainty the phalanges of the
fore from those of the hind foot, they will be noticed after the description of the meta-
tarsals.

The Metatarsals of Hyopotamus and Diplopus.

My materials for the hind foot are fortunately more complete than those for the fore
limb, and the striking dissimilarity we noticed between Diplopus and Hyopotamus in
considering the manus, is still more confirmed by the study of the pes. According to
the order adopted by me, I shall describe at first the middle metatarsals, as they always
exhibit the fundamental features of the pes, and consider the lateral digits afterwards.
I begin with the tetradactyle Hyopotamus,

The third metatarsal (Plate XXXVIII. fig. 1, in.). — As in all Paridigitata, this meta
tarsal is distinguished at once by the presence, on its fibular side, of a deep pit, into
which fits the corresponding projection of the tibial side of the fourth (fig. 1, iv.),
giving rise to the characteristic interlocking of the middle digits. The proximal
surface of the third metatarsal (fig. 2', in.) is of a rounded triangular outline, and
slightly concave to meet the slight convexity of the distal surface of the third cunei-
form (fig. 1, c 3 ). The third metatarsal is supported entirely by the third cuneiform ;
and as the lateral internal (tibial) side of this third cuneiform descends a little lower
than the second cuneiform, the second metatarsal going to meet its typical second
cuneiform, may lean also against the third cuneiform (fig. 2, n.), a constant feature in all
unreduced Ungulata. The posterior projection of the metatarsals is very long, and pressed
against a similar projection of the fourth metatarsal (fig. 2', in., iv.) ; on the inner (tibial)
side of this projection we see no trace of a facet for the first cuneiform, such as is
presented by Diplopus (fig. 4-^fc^ and the recent Suina. The inner or tibial edge of the
proximal surface is slightly elevated, though not enough to exclude the second meta-
tarsal from its facet on the third cuneiform, as it may be seen also in Hippopotamus
and Anthracotherium. The outer, or fibular, side of the third metatarsal is flattened
(section, fig. 1) in correspondence with the adjoining side of the fourth; besides the
deep pit mentioned before, we see, on this side, a long narrow facet on the posterior
projection (fig. 2'), articulated to a similar facet of the posterior projection of the
fourth. (As the fourth metatarsal was slightly defective in my specimen, the posterior

OSTEOLOGY OF THE HYOPOTAMIM2. 85

projections of the two middle digits that must in reality articulate together are sepa-
rated in the drawing.) The inner or tibial side of the third metatarsal shows an oval
facet for the articulation of the second digit ; a like facet is seen in the Suina and
Hippopotamus, only much shorter.

The fourth metatarsal (Plate XXXVIII. fig. 1, iv.), — What at once strikes the
observer on looking at this metatarsal, is the large projection of its tibial side, which
enters into the corresponding pit of the third metatarsal ; a little below it the inner
surface of the bone is very rugose and bulging a little inwards. The proximal surface
(fig. 2', iv.) is slightly concave at its fore and inner part, and somewhat raised in the
postero-external angle (the elevation is indicated in fig. 2' by a deeper tint) ; this raised
joint fits exactly into the postero-external concavity of the distal surface of the cuboid
(fig. 10, iv.), while the remaining, and slightly convex, cuboidal surface is fitted to the
slight concavity of the inner and fore part of the proximal surface of the fourth meta-
tarsal.

The outer side of the fourth metatarsal has a lengthened oval facet for the articula-
tion of the fifth digit ; such a facet is to be seen in the Hog, only a little shorter.
Whether the fifth digit articulated with the fourth by a second facet I am unable to say,
as my specimen is a little defective ; but very probably it did.

As the two middle digits, in their general shape, bear a great likeness to each other,
all I shall say of one will be referable to the other. Their symmetry is somewhat
disturbed by the slight bulging of the inner side of the fourth metatarsal; although,
if we look at the anterior surface of the whole pes, this slight disturbance does not
interfere with the general symmetry of the two middle digits.

The section of the middle digits (Plate XXXVIII. below, fig. 1) has a flattened tra-
pezoid outline, especially if we take it in the upper part, where the posterior projection
is prolonged downwards as a flattened platform in the upper half of the posterior surface
of the metatarsals; towards the middle this platform subsides, and we have a more
rectangular section. This flatness of the metatarsals is very striking in comparison with
the rounded outline of the metatarsals in the didactyle genus ; but one of the living
Paridigitata, the Hippopotamus, has even much flatter metatarsals*, their thickness
being only half of their transverse breadth. The outer margins of both middle meta-
tarsals are made conspicuous and angular by the pressure of the completely developed
lateral digits; their anterior surface is therefore very flat, even more so than in the living
Hippopotamus. The middle metatarsals preserve a uniform breadth along their entire
length, and we see no such conspicuous broadenings of the distal ends as in the didactyle
form. The restored pes (Plate XXXVII. fig. 21) does not show this very clearly, as the
complete digits from Puy were much disfigured by pressure ; but it could be readily seen
on some well-preserved fragments of the distal extremities. The distal end is quite

* I find in a Hippopotamus with epiphysed bones that the fourth and third metatarsal have 32 mm. trans-
verse breadth each, and 16 and 17 mm. thickness or depth, while in Hyopotamus the breadth is 16 each and the
depth 11 mm.

N 2

86 DE. W. KOWALEYSKY ON THE

smooth anteriorly, the articular ridge for the first phalanx being limited to the posterior
or palmar half of the distal surface.

The lateral digits were completely developed in Hyopotamus and, no doubt, played
an active part in the process of locomotion. The second metatarsal, seen from the
inner side in Plate XXXV1IL fig. 2, n., is articulated to the second cuneiform by a
nearly circular, flat proximal facet; this facet occupies the whole upper, somewhat
attenuated, head of the second metatarsal ; the anterior part of this proximal head is
pressed against the third cuneiform, as seen in fig. 2, n. c 9 . On its postero-tibial side,
a little lower down, is seen another elongated facet, to which was articulated the first
cuneiform, although this last bone was not to be found in any of the collections I
visited. The shaft of the second metatarsal was closely pressed against the third, and
reached very low down the metacarpus, considerably lower than in the Hog; and in
my restoration of the pes (Plate XXXVII. fig. 21) the lateral digits, as I am aware now,
are too much shortened. The section of the second digit is perhaps more elongated than
in my figure. The distal extremity is unsymmetrical, but very well developed, the ridge
for the first lateral phalanx being very high and confined to the back part of the distal
end. The truncated posterior border of the proximal end of the second metatarsal is
very characteristic of all Paridigitata which have retained the lateral digits : it is to be met
with in the Suina, in Cainotherium, and even in Hippopotamus ; this truncature is
intended for the first cuneiform, which articulates with the navicular, with the posterior
part of the small second cuneiform, and, by a large facet, with the truncated posterior
edge of the second metatarsal *.

The fifth (Plate XXXVIII. fig. 1, v., Plate XXXVII. fig. 21, v.) or outer metatarsal
presents, at its proximal end, a triangular facet, corresponding to the facet v. on the
distal side of the cuboid (Plate XXXVIII. fig. 10) ; the posterior end of this fifth
digit is prolonged backwards into a projection which has very nearly the same antero-
posterior length as the fore or articular part. The inner or tibial side has a half
reniform facet for the articulation with the fourth metatarsal. The section of this
digit in the middle gives a somewhat roundish outline ; its general shape is oval, the
proximal third is a little curved forwards to fit more exactly the outline of the neigh-
bouring fourth metatarsal.

The Metatarsals of Diplopus.

Of these I found several specimens in a perfect state of preservation in the
collection of the British Museum f . They are all said to come from Hordwell ;

* In Hyopotamus, although the third cuneiform has no regular truncated edge, as in Hippopotamus (Plate
XXXVII. fig. 10, c 3 ) and Anthracotherium, for the articulation of the second metatarsal, still this metacarpal,
owing to the fact that the second cuneiform is situated a little higher than the third, is allowed to touch this
last, remaining thus true to typical relations.

f On my second visit to Puy, after this paper was written, I saw, in the collection of Mr. Vi;n t ay, a detached
second metatarsal of unusually large size. It belonged to the large species, and, judging by this bone, all the
four metatarsals of the larger. species of Hyopotami were nearly subequal, as in the recent Hippopotami®.

OSTEOLOGY OF THE HYOPOTAMIM. 87

and I must state that I have not seen a single specimen belonging to the didactyle
JDiplopus from Hempstead, and, vice versa, not a single specimen of the tetradactyle
Hyopotamus from Hordwell. The state of preservation of the fossils in both deposits
is very different, and, as far as I can judge from information received in the
British Museum, the tetradactyle genus Hyopotamus, identical with the genus from
Puy, seems to be confined to Hempstead, while the didactyle form is found only at
Hordwell *. I hope, however, to discuss the stratigraphical questions more fully at the
end of my paper.

The two middle metatarsals of Diplopus, seemingly from the same individual, with
three tarsal bones, are figured of the natural size in Plate XXXVIII. figs. 3 & 4, and
two others, from a larger individual, Plate XXXV. fig. 5. The!three phalanges of the
fourth digit are a restoration, as I had only the corresponding phalanges from the other
side ; the two metatarsals are also from different individuals, and the third is a little
smaller than the fourth.

The difference of these two middle digits of the pes from the corresponding bones of
Ilyopotamus is very great in general shape, section, and in some of the minor relations
to the tarsal bones, the principal relations being the same in both genera.

The dissimilarity in shape is of the same order as that noticed in reference to the meta-
carpals. Instead of the flattened and angular metatarsals of Hyojpotamus as seen in
section Plate XXXVIII. fig. 1, we have, in Dijplopus, very rounded, mutually symme-
trical digits, which adapt themselves mutually by a large, flat inner side, and are sym-
metrically rounded both in and outside, so as to present in section nearly a perfect half
cylinder (Plate XXXVIII. v., figs. 3, 4, 5). By this shape they differ much from the middle
digits of the equally didactyle Anoplotherium and Anthracothermm magnum (the digits
of which I have from Pochette, Lausanne), but present a resemblance to the same bones
of the Xiphodon and Entelodon (didactyle), and also to that of the common Hog, as this
last may be said to be practically didactyle. As seen in Plate XXXV. fig. 5 the distal
ends are considerably broadened, and the articular ridge, though confined to the palmar
side of the distal extremity, is prolonged in the form of a very low, but visible, elevation
along the lower end to the anterior surface of the distal articulation (see Plate XXXVIII.
fig. 6, and Plate XXXV. fig. 5). These slight peculiarities, which are too numerous
to be all noticed, clearly indicate a somewhat better adaptation to didactyle locomotion,
or at least some nearer approach to our recent didactyle forms than is exhibited by older
forms, such as Anoplotherium — every new experiment of nature to produce a didactyle
genus seeming to be more successful than the preceding one. However, I have no doubt
the experiment ended there; and no direct connexion exists between the didactyle

* I have seen a very large lunare from Hempstead, which appears to indicate the presence of a larger
Ifyopotamus in this deposit ; a very large cuboid from the same locality, though much rolled, seems to indi-
cate, so far as can be judged by its imperfect state, the presence of another very large tetradactyle species at
Hempstead. The cuboid in question presented on its distal surface two facets for the fourth and fifth digits,
and no beak-like posterior projection.

88 DE. W. KOWALEVSKY ON THE

Hyopotamidee and the living genera of Paridigitates, which, as it seems to me, have
descended from a branch given off by the tetradactyle Hyopotamidae of the Eocene epoch.

The proximal articular surfaces of the two middle metatarsals of Diplopus (Plate
XXXVIII. fig. 4") may be compared with the corresponding proximal surfaces of the
tetradactyle Eyopotamus (Plate XXXVIII. fig. 2'), when the differences are seen at a
glance. The facet (cb.f) is intended to meet a like facet on the inner side of the beak-
like posterior projection of the cuboid (fig. 11, f. my) ; it is wanting in Hyopotarrms, as
well as the cuneiform facet (fig. 4",/^) for the first cuneiform, which was wedged in
between this posterior projection of the third metatarsal and the rudiment of the second
digit (fig. 4 /; , nr ). This rudiment was confluent with the third metatarsal in one of
the figured specimens (fig. 4 and 4", nr), but it was absent (because not so coalesced)
from all the other specimens of the third metatarsals in the British-Museum collection- —
for instance, from the third metatarsal represented in Plate XXXVIII. fig. 12. On the
free (outer and inner) sides of both metatarsals are longitudinal facets to which the
rudiments were articulated ; these last, however, I could not find, save the one of the
second metatarsal ankylosed to the third represented in fig. 4, Plate XXXVIII.

The interlocking of the two middle metatarsals was effected by a very prominent tubercle
of the fourth entering a deep pit on the fibular side of the third, as seen in the figures ;
besides there is an oval facet on the inner faces of the posterior projection of each meta-
tarsal. The beak-like downward process of the cuboid and the wedge of the first cunei-
form pressing laterally from the inner and outer sides upon these posterior projections of
the two middle digits held them firmly together. This close fitting of the two metatarsals
(and metacarpals) together was further assisted by the mode of metatarsophalangeal
articulations universal among Paridigitata — namely, by the outer halves of the distal
surface being a little shorter than the inner (as is clearly seen in Plate XXXVIII.
fig. 6) : in the phalanges the relation is inverse, and therefore in treading on the ground
their upper ends are made to converge and to press the two metatarsals and metacarpals
together ; the two separate metatarsals, compressed in this way, approach as near as
possible to the cannonbone of modern Euminantia.

Phalanges.

The first phalanges. — In a set of mixed phalanges belonging to both Diplojms and
Eyopotamus it would be utterly impossible to distinguish the bones belonging to each
genus : as difficult is it to separate those of the manus from the phalanges of the pes ;
their relative height and thickness present no good constant characters. One of the
best authorities, Professor Hensel, in his memoir on Hipparion, printed in the Transac-
tions of the Berlin Academy for 1861, tells us that it is even impossible to distinguish
the fore and aft phalanges of living Ungulata in the case of several individuals being
mixed together. This may serve as an excuse for my not trying to do it among the
fossils ; and I therefore intend to give only a general description, which will apply to
both the fore and hind limb.

OSTEOLOGY OF THE HYOPOTAMID,£. 89

The phalanges of Hyopotamus found at Puy, where the didaetyle Diplopus is not
present, are, by this fact, already separated by nature. As seen in Plate XXXVII.
figs. 20 & 21, they strike us as being much longer than in most of the living genera,
and considerably longer than in Anoplotherium. In Xiphodon and Entelodon, however,
the first phalanges are also very long.

The upper, or proximal, articular surface is not symmetrical, as the inner side of the
first phalanx is considerably thicker than the outer, in correspondence with the same
inequality of the distal extremities of the metacarpal and metatarsal bones. The same
difference in the thickness of the inner side is also to be seen on the distal extremity.
The proximal surface is concave, and the groove for the articular ridge of the metacarpal
is limited to its posterior third only. This thickening of the inner side makes also the
inferior end not completely symmetrical ; and this want of symmetry is much greater in
the first phalanges of the didaetyle Diplopus.

The second phalanges. — The second phalanges of all Paridigitata (with the exception
of Hippopotamus and Camelidce) are very characteristic, as their distal extremity is shaped
unsymmetrically in a peculiar manner, so as to cause the ungual phalanges to converge in
treading on the ground. For this purpose, the outer half of the distal articular surface is
not only much larger than the inner, but bends obliquely inwards, and the ungual pha-
langes following this inner curve tend to converge. This arrangement is very strongly
developed in the second phalanges of Anoplotherium (see De Blainville, Osteogr.
Anoploth. pi. iii.) ; it may be seen in every Euminant as well as in the Suidse. Hyopotamus
and Dvplopus also have second phalanges shaped on this pattern ; only this want of sym-
metry is not so clearly developed as in the Suina or Euminantia.

The third phalanges. — These have a very peculiar shape and are quite identical in both
genera. This strange shape I can compare to nothing better than to a very thickened
and rounded human nail. Their proximal surfaces are unsymmetrical, to fit the unsym-
metrical distal end of the second phalanges ; but the remaining part is much more symme-
trical than in the Suina or Euminantia, the inner side not being flattened at all, or very
slightly. The palmar surface is quite flat. The lower and anterior margin shows the
usual vascular foramina and a certain crispness to allow a firmer fitting of the horny hoof.
With this I conclude my description of the long bones of the skeleton and the bones
of the limbs ; the latter are all figured of the natural size, and the sections give a correct
idea of their breadth and antero-posterior depth. Some exact dimensions which could
be taken are given in the general Table (p. 90) I have taken the liberty of disposing
right and left as best suited my purpose; and while some bones from the collection
of the British Museum were drawn directly from nature, and therefore in the Plates
appear to belong to the opposite side, others have been reversed.

DB. "W. KOWALEVSKY OJSf THE

Table of Measurements of the Extremities *.

Hyopoiamus,

Metacarpals.

Length

Transverse breadth, proximal face

„ „ middle

„ „ distal extremity

Antero-posterior depth, in the middle

Metatarsals.

Length

Transverse breadth, proximal face ..........

„ „ middle ................

„ „ distal extremity

Antero-posterior depth in the middle

Eirst Phalanges.

Length . ;

Transverse breadth, proximal face ..........

„ „ distal extremity

Second Phalanges.

Length

Transverse breadth, proximal face ...

„ „ distal extremity ,

Third Phalanges.

Length

Transverse breadth, proximal face ..........

Puy.

105

101

113

18
16

• # • «

IV.

i«4

15J

• • • •

48, 49, 50
15, 16, 16^

16, 15|
14, 13^

27,28

JLTtj JL JLo"

Hempstead.

III.

IV.

16
15

# • • •

14, 13
9,10

37, 38, 40

16, 16, 17

• * • »

* • • *

Diplopus.

Hordwell.

JL-L-L*

• * * «

• • • a

• * • •

140
18,15
15, 14
20
16,15

IV.

138

20
18
22
16

156
22, 18
16,17

w'rfco"

17,16

51
25
171

25,30
18,19
16, 17

29
15

Explanation of the Plates.
PLATE XXXV.

Fig. 1. Scapula of Diplopus, Hordwell, f nat. size. Brit. Mus.

$, acromion ; cp, coracoid process.
Eig. 2. Distal extremity of the femur oiHyopotamus, Hempstead. Mus. Pract. Geology.

it, internal trochlea.
Fig. 3. Extremity of the fibula of Diplopus, nat. size, Hordwell. B.M.
Fig. 4. Calcaneum of Diplopus, Hordwell. B.M.

//, fibular facet ; as, astragalean facet ; cb, cuboid facet.

* The absence of a number indicates that the bone was broken at that particular place and did not allow of
exact measurement. My materials for the lateral digits were much more scarce than for the middle ones, as
the former, through their slenderness, are often broken and lost. Nearly all the bones of Hyopotamus figured,
from Puy and Hempstead, belong to the small species Hyopotamus velaunus, Aym. Such measurements as
I could take of the lateral digits show that they have half the breadth of the middle ones. So the breadth of
the second metacarpal from Puy is 8 millims., the depth 10 millims. • the second metatarsal is 8| millims.
broad, and 6 deep at the attenuated proximal end ; the fifth metacarpal is 9=§ millims. broad and 8 millims.
deep at the proximal end ; the fifth metatarsal is 91 millims. broad and 9 millims. deep.

OSTEOLOGY OF THE HTOPOTAMIDJE. 91

Fig. 4'. Calcaneum of a Perissodaetyle (Anchitherium).

i, internal astr. facet ; 0, external astrag. facet; s 9 sulcus sustentaculi.
Fig. 5. Two metatarsals of Diplopus, left foot, from the New Forest (Brockenhurst).

B.M. The phalanges of the third digit are restored.
Fig. 5'. Section of these two metatarsals in the middle.

All the figures except the scapula are of the natural size.

PLATE XXXVI.

Figs. 2 & 3 are of the natural size, all others two thirds of the natural size.

Fig. 1, side, and fig. 1/, front view of the ulna of Diplopus from Hordwell. B.M.
or, outer radial facet ; ir, internal radial facet.

Fig. 2. Front view of the upper part of the ulna oiHyopotamus, Puy : or, external radial
facet ; ir, internal radial facet ; cb> connecting bridge.

Figs. 1" & 2'. Sections of both ulnee : e, external ; i, internal edge ; a, anterior ; jp, pos-
terior surface.

Fig. 3. Proximal part of radius of Hyopotarnus, Hempstead. B.M.

Fig. 3'. Distal extremity of the radius of Hyopotamus from Hempstead.
ex, external ; i 9 internal side, B.M.

Fig. 4. Humerus of Diplopus, from Hordwell. B.M.

<?, intercondylar perforation; a, middle bulging ; b, internal projection.

Fig. 4'. Section of the same.

Fig. 5. Femur of Hyopotamus from Puy.

t.mj, trochanter major; t.mn, trochanter minor.

Fig. 6. Smaller femur from Hempstead. B.M.

Fig. 7. Tibia and fibula of Diplopus, Hordwell. B.M.

Fig. 7'. Distal view with the fibula. Fig. 7". Section about the middle of the bone.

PLATE XXXVII.

Letters common to all the figures : — s, scaphoid ; I, lunar ; p, pyramidal ; tz 9 trapezium ;
t, trapezoid ; m, os magnum ; u, unciform ; c, calcaneum ; a, astragalus ; n, navi-
cular ; cb, cuboid ; <? 3 , third, c 2 , second, and c u first cuneiform.
Fig. 1. Right fore foot of Hippopotamus; fig. 2 of Anoplotherium tridactylum from

Vaucluse; fig. 3 of Xiphodon.
Fig. 4. Side view of the right fore foot of the Hog; fig. 5 of Dicotyles; fig. 7 of

Tragulus Kantchil ; fig. 8 of HyomoscJius aguaticus.
Fig. 6. Third right metacarpal of Palceochoerus (Allier j, to show the absence of the radial
enlargement.

MDCCCLXXI1I. N*

92 DB. W. KOWALEVSKY ON THE

Fig. 9. Right hind foot of Hippopotamus. Fig. 10. Side view of the same, to show
the articulation of the third and second metatarsals with the first, second, and
third cuneiforms.

Fig. 11. Left hind foot of Anoplotherium tridactylum.

Fig. 12. Side view of the right hind foot of the Hog; articulation of the third and
second metatarsals with the three cuneiforms. Fig. 12'. Third metatarsal of
Palcebehoerus. Fig. 13. JDicotyles. Fig. 14. Hyomosehus aquaticus. Fig. 15.
Tragulus Kantchil. Fig. 16. Amphitragulus (1) from Allier.

Fig. 17. Distal surface of the navicular of Anoplotherium commune, Paris gypsum, to
show the facets for the three cuneiforms.

Fig. 18. The same bone from Vaucluse. Fig. 19. The same bone, Anoplothermm tri-
dactylum, to show the great development of the facet for the second cuneiform
(c 2 ), which carries the second digit.

Fig. 20. Fore foot of Hyopotamus from Puy (partly restored).

Fig. 21. Hind foot of the same.

Figs. 1, 9, and 10, \ nat. size; figs. 2, 11, 20, and 21, \ nat. size; all others of the
natural size.

PLATE XXXVIII.

Same letters as in Plate XXXVII,

Fig. 1* Hind foot of Hyopotamus, Puy.

as, astragalean ; el, calcaneal facet of the cuboid.
Fig. 2. Same, side view. The third and second cuneiforms are coalesced ; the first
cuneiform is absent.
Below fig. 2, proximal faces of the four metatarsals arid section of the same.
Fig. 3. Hind foot of Diplopus, Hordwell. B.M.
Fig. 4. Side view of the same.

bxb, beak of the cuboid; fo 19 facet for the first cuneiform; iir, rudiment of the

second metatarsal coalesced with the third.
Fig. 4'. Section of the two middle metatarsals at half their length.
Fig. 4 /; . Proximal view of the .metatarsus of Diplopus.

ch.f, facet for the descending beak of the cuboid; fc x , facet for the first cunei«

form; nr, rudiment of the second metatarsal coalesced with the third.
Outline of the distal face of the two coalesced cuneiforms (third and second) of
Diplopus, fitting the proximal face of the third metatarsal with the coalesced
rudiment of the second metatarsal (nr). B.M.
Fig. 5. Fore foot of Hyopotamus, Puy.
Fig. 6. Two metacarpals of the fore foot of Diplopus, Hordwell. B.M.

OSTEOLOGY OF THE HYOPOTAMDXE. 93

Fig. 7. Distal view of the unciform of Hyopotamus from Puy.

m, iv, v ? facets for the corresponding metacarpals.
Figs. 8 & 8'. Front and distal view of the unciform of Diplopus, HordwelL Mus. Cam-
bridge.
In, lunar facet ; W , pyramidal facet ; in, v, facets for the two metacarpals.
Fig. 9. An unciform of Hyopotamus from Hempstead. Fig. 9'. Distal view of the same.

Smaller than the unciform from Puy. B.M.
Fig. 10. Distal view of the cuboid of Hyopotamus, Puy.

tr, transverse ridge ; iv, v, facets for the two corresponding metatarsals.
Fig. 11. Distal view of the cuboid of Diplopus. B.M.

f.my, facet for the fourth metatarsal on the posterior beak ; iv, facet for the
single fourth metatarsal.
Fig. 12. Another third metatarsal of Diplopus, without the rudiment of the second,,

All the hind feet are right ; the fore foot is left

t>t a 'TT? "VV^VTUT

Fig. 1. Cranium of the largest species of Hyopotamus, from Puy. Two thirds nat

size.
Fig. 2. Side view of the head of a smaller Hyopotamus, from Puy. Nat. size.
Fig. 3. Lower jaw of the same ; said to come from the same block.
Fig. 3'. Upper view of the anterior extremity of the lower jaw.
Fig. 4. One molar (m l ) and three milk-molars of Hyopotamus, Puy.
Fig. 5. Part of the head, from the postglenoid process to the occipital condyle.
Fig. 6. Dorsal vertebra of Hyopotamus.

Fig. 7. Second cervical vertebra of Hyopotamus, from Puy. Collection Pichot.
Fig. 8. Left lower molar of Chalicotherium. Fig. 9. Anoplotherium. Fig. 10. Dicho-

bum bavarica, Fraas. Fig. 11. Hyopotamus. Fig. 12. Euminantfrom Allier.
ac, anterior crescent ; pe, posterior crescent ,* ap, anterior pillar ; pp, posterior

pillar.
Fig. 13, Left upper molar of Bhagatherium ; fig. 14 of Hyopotamus Gresslyi. Fig.

15. Intermediate form between the Hyopotamus and Diehodon ; the anterior

middle (fifth) lobe is coalesced with the internal. Fig. 16. Diehodon, from

Mauremont.
el, external anterior lobe; ml, middle anterior lobe; il> internal anterior lobe.
The teeth (figs. 13, 14, 15) have the three lobes on the fore part of the tooth.
Fig. 17. Left upper molar of Dichobune leporina, Cuv. ; and fig. 18 of Cainotherium:

the three lobes are on the posterior part of the tooth.
Fig. 18. The so-called Cainotherium (Hyopotamus) Benevieri, Pict., having, like all the

other Hyopotamidw, the three lobes on the anterior part of the tooth.

94 DE. W. KOWALEVSKY ON THE OSTEOLOGY OE THE HYOPOTAMID^E.

t>t AHPT? YT
JTXjxxXJQi jOlju,

Fig. 1. Head of Hyopotamm velawms, Aym., from Puy ; restored, bat belonging to one
individual : all the pieces were found together in the same block. Side view.

Fig. 2. View from above, to show the great sagittal crista.

Fig. 3. Lower jaw of a very young Hyopotamus from Puy ; the first molar is concealed
in the jaw.
d\ d 2 , d 3 , the three milk-molars ; $>*, the first premolar which has no milk-tooth
to precede it.

Fig. 4. Upper view of the three milk-molars,

JKn-waLev shy.

Fhub: I-rcurv, MDCCCLXXIIL FlxuUJUSf.

>dw~ cp

C .L. Qrie sbaLcTi Mh .,

llTntern. Bros, imp.

JKow aJ.xi-vs hv

ThzL. Trans. MDCCCLXXT1I Plate XX1VL

CXi . Oris sb&ch. litL

"Minfcerii-Bros. imp.

KowaLwsky.

PhiJb. Trans. MDCCCLXXIH, PlcO&HWTL.

_.iz

G .L . Grio sl)ack del. et lith.

Minterri. Bros, imp.

J&waZ&vsky.

Phul. Trans. MDCCCLXXttl. PlcuUXXXVKl

^~Jo.cb

BLrnXy,

C.L .Griesba£L -lift..

IVO-iitem Bros. imp.

huh'.'J IfV^n'V

,s*

:<^^

««,-

'Phil. Trans. MDCCCmill. Pi a.u- XXXIX.

S"' "<

•M:".fi««?.v<

1P ^ -

M'.t's-'I'Ui'h dyl et lit K.

^SKSS

',**"»--»i<6?'* > '

<7 7

.---<ii^' ;

,.ijjk*»*v** ii *

VJ*

V.v

jtfp^ig&r. ■*

?* '■'

Mmtei-a iirc,

. ;tr;p.

Kowal&vsky.

TJvlZ. Trams. MDCCCKXXIIL JPZoataXL

.^r^-^-^^^S^^

G'li. GitiesbajciL lith.,

Mmterri'Bros. imp.